Author + information
- Received August 4, 1999
- Revision received November 9, 1999
- Accepted December 29, 1999
- Published online April 1, 2000.
- Arne SippensGroenewegen, MD∗,†,* (, )
- Michael D Lesh, MD, FACC∗,†,
- Franz X Roithinger, MD∗,†,
- Willard S Ellis, PhD∗,†,
- Paul R Steiner, MD∗,
- Leslie A Saxon, MD, FACC∗,
- Randall J Lee, MD, FACC∗,† and
- Melvin M Scheinman, MD, FACC∗
- ↵*Reprint requests and correspondence: Dr. Arne SippensGroenewegen, Section of Cardiac Electrophysiology, Department of Medicine, University of California at San Francisco, 500 Parnassus Avenue, MU-East 4S, Box 1354, San Francisco, California 94143-1354
Data included in this study were, in part, presented at the 70th Scientific Sessions of the American Heart Association, November 8–12, 1997, Orlando, Florida and the 19th Scientific Sessions of the North American Society of Pacing and Electrophysiology, May 6–9, 1998, San Diego, California.
This study was directed at developing spatial 62-lead electrocardiogram (ECG) criteria for classification of counterclockwise (CCW) and clockwise (CW) typical atrial flutter (Fl) in patients with and without structural heart disease.
Electrocardiographic classification of CCW and CW typical atrial Fl is frequently hampered by inaccurate and inconclusive scalar waveform analysis of the 12-lead ECG.
Electrocardiogram signals from 62 torso sites and multisite endocardial recordings were obtained during CCW typical atrial Fl (12 patients), CW typical Fl (3 patients), both forms of typical Fl (4 patients) and CCW typical and atypical atrial Fl (1 patient). All the Fl wave episodes were divided into two or three successive time periods showing stable potential distributions from which integral maps were computed.
The initial, intermediate and terminal CCW Fl wave map patterns coincided with: 1) caudocranial activation of the right atrial septum and proximal-to-distal coronary sinus activation, 2) craniocaudal activation of the right atrial free wall, and 3) activation of the lateral part of the subeustachian isthmus, respectively. The initial, intermediate and terminal CW Fl wave map patterns corresponded with : 1) craniocaudal right atrial septal activation, 2) activation of the subeustachian isthmus and proximal-to-distal coronary sinus activation, and 3) caudocranial right atrial free wall activation, respectively. A reference set of typical CCW and CW mean integral maps of the three successive Fl wave periods was computed after establishing a high degree of quantitative interpatient integral map pattern correspondence irrespective of the presence or absence of organic heart disease.
The 62-lead ECG of CCW and CW typical atrial Fl in man is characterized by a stereotypical spatial voltage distribution that can be directly related to the underlying activation sequence and is highly specific to the direction of Fl wave rotation. The mean CCW and CW Fl wave integral maps present a unique reference set for improved clinical detection and classification of typical atrial Fl.
Clinical electrocardiography of typical atrial flutter (Fl) dates back to the early part of this century when the continuous “sawtooth” pattern of the Fl waves in leads II and III was first reported (1,2) and attributed to an atrial impulse circulating around both caval veins (3). Later reports provided further evidence to the concept of an excitation wave rotating in a counterclockwise (CCW) direction in the right atrium and giving rise to the distinct “sawtooth” pattern with predominantly negative Fl waves in the inferior leads and V6 combined with a positive Fl wave in V1(4,5). With the introduction of multisite endocardial mapping (6–13), entrainment techniques (6–15) and catheter ablative strategies in the therapeutic management of the typical form of atrial Fl (8–11,13,16,17), Fl wave propagation has been convincingly demonstrated to occur in a macroreentrant right atrial circuit confined anteriorly by the tricuspid annulus and posteriorly by the crista terminalis and its inferior extension, the eustachian ridge, which act as natural barriers of conduction (10,15). In addition to CCW rotation of the Fl wave, it has been shown that clockwise (CW) impulse rotation in the same right atrial circuit can be frequently observed (6,11–13,17). Although there is general concensus on the aforementioned distinct morphology of the 12-lead electrocardiogram (ECG) during CCW typical atrial Fl (7,8,10–13,17), electrocardiographic features found to be most specific for CW typical atrial Fl were more variable and have included a predominantly positive Fl wave polarity in the inferior leads (6,12,17) and, additionally, V6(13) as well as a short plateau phase with a wide negative component in the inferior leads and an overall negative Fl wave in V1(11). Visual assessment of the Fl wave polarity on the 12-lead ECG is hampered by low-voltage, a continuously undulating signal (6,12) and QRST wave obscurement when a low degree of atrioventricular (AV) block is present (11). Moreover, the 12-lead ECG has been found to be insufficient in distinguishing between CCW and CW typical atrial Fl (11) or between the two forms of typical Fl and atrial Fl not dependent on the subeustachian isthmus (13). Given our recent favorable clinical experience with the use of body surface mapping in providing detailed spatial localization of ectopic right atrial activation (18), we felt that the application of this technique might improve the performance of the surface ECG in characterizing and classifying typical atrial Fl (19).
This study features the application of 62-lead ECG mapping in conjunction with multisite endocardial catheter mapping in patients with typical atrial Fl to: 1) document the instant-by-instant changes of the CCW and CW Fl wave patterns on the body surface, 2) compare the body surface Fl wave patterns with the underlying site of endocardial activation within the reentrant circuit, and 3) develop a reference set of body surface map patterns characteristic for CCW and CW Fl wave rotation.
A total of 20 consecutive patients with recurrent typical atrial Fl referred for catheter ablation were included in the study. There were 17 men and 3 women (mean age 66 ± 8 years) of whom 13 patients (65%) had concomitant structural heart disease (Table 1). None of the enrolled patients had undergone prior ablation of typical atrial Fl or atrial surgery for congenital heart disease. Five patients (25%) were receiving ongoing treatment with low-dose amiodarone or had stopped taking this drug one to three weeks before the study. All other antiarrhythmic agents were discontinued at least five drug half-lives before the study. Beta-adrenergic blocking agents or digitalis therapy was administered in seven patients (54%). Informed consent was obtained from all patients in accordance with the protocol approved by the Committee on Human Research of the University of California at San Francisco.
Multisite endocardial mapping
Four multipolar catheters were introduced percutaneously and positioned using biplane fluoroscopic monitoring in the 45° right anterior oblique (RAO) and 45° left anterior oblique (LAO) projections. A 7F 20-pole “Halo” catheter (2-10-2 mm interelectrode spacing) (Cordis-Webster, Inc., Diamond Bar, California) was placed along the tricuspid annulus so that the distal electrode pair was located at the lateral entrance of the subeustachian isthmus (7:00 o’clock on LAO view) and the proximal electrode pair was situated at the high septum (2:00 o’clock on LAO view). A 6F octapolar catheter (2.5 or 2.5–5-2.5 mm interelectrode spacing) (EP Technologies, Inc.) was positioned at the His bundle position or midseptum, while a 7F open-lumen decapolar catheter (5 mm interelectrode spacing) (DAIG, Inc., Minnetonka, Minnesota) was advanced into the coronary sinus, thereby ensuring that the proximal electrode pair was situated at the os of the coronary sinus. When necessary, the exact position of the latter catheter was verified fluoroscopically using injection of contrast dye. Finally, a 7F roving quadripolar catheter (2.5 mm interelectrode spacing) with a 4 mm tip electrode (EP Technologies, Inc., Sunnyvale, California) was used for activation and entrainment mapping in the subeustachian isthmus.
Bipolar endocardial electrograms and the 12-lead surface ECG were acquired, processed and analyzed using a 48-channel CardioLab system (Prucka Engineering, Inc., Houston, Texas). The signals were filtered at a bandpass setting of 30 to 500 Hz, digitized at a 1 kHz sampling rate and stored on optical disk. The signals were displayed on a color monitor or produced as hard copy at variable gain settings with a paper speed of 200 to 400 mm/s.
Arrhythmia induction was obtained with programmed atrial stimulation when sustained atrial Fl (>30 s duration) was not present at the onset of the mapping procedure. Verification of CCW or CW typical atrial Fl was based on the characteristic sequence of activation and the demonstration of concealed entrainment with pacing in the subeustachian isthmus (6–14). Entrainment mapping in the subeustachian isthmus was carried out according to previously described techniques (10,15).
Body surface mapping
Recording and processing of electrocardiographic data was performed using an on-line portable body surface mapping system (20,21) and a radiotransparent carbon electrode array according to recently published methods (18). Briefly, unipolar ECG signals of single heart beats were acquired simultaneously from 62 lead sites (Fig. 1A) using Wilson’s central terminal as a reference. Lead sites in the array are referred to with a letter (A to P) and numeric code (1 to 12) according to a division in 16 columns and 12 rows, respectively; this means that electrode “A1” is located at the right anterior axillary line just below the level of the suprasternal notch. Recordings of the surface ECG and endocardial catheter mapping of spontaneous or electrically induced sustained atrial Fl were performed simultaneously during resting tidal volume respiration. Whenever a low degree of AV block occurred (i.e., ≤2:1 AV ratio) resulting in a partly obscured Fl wave due to the superimposed QRST segment, adequate isolation of the complete Fl wave cycle was obtained using intravenous delivery of adenosine (Table 1) (11). The ECG signals were amplified, digitized at a 1 kHz sampling rate using a 14-bits A/D converter, optically transmitted and stored on hard disk as well as removable hard disk cartridges. All lead tracings were visually screened in order to reject poor quality signals or signals from lead sites that were obscured by defibrillation patches on the front and back of the chest (mean 4.2 ± 4.3 per map). Replacement values were interpolated based on adjoining lead recordings. Digital filtering to remove line frequency interference was performed when necessary with a 60-Hz notch filter. Correction of interelectrode offset differences and linear baseline drifting was performed using linear interpolation after selecting an isoelectric time instant. Due to the absence of a true isoelectric period in the Fl wave cycle, an interval was selected in the signal near to the zero voltage level directly preceding the negative inscription of the Fl wave (19). A series of potential maps of the entire Fl wave cycle was computed at 2-ms intervals. The potential maps were visually evaluated to determine Fl wave onset and offset as the point in time when one of the extreme values superseded a ±30 μV voltage window around the zero line (18) and all extreme values had returned within this ±30 μV voltage window, respectively. Integral maps of selected time intervals in the Fl wave cycle were computed for each recording (please refer to the “Data Analysis and Statistics” section for further details) (Fig. 1B).
Counterclockwise typical atrial Fl was defined as a right atrial arrhythmia characterized by CCW wavefront rotation around the tricuspid annulus and critical dependence on the subeustachian isthmus (8–10,13). Clockwise typical atrial Fl was defined as as a right atrial arrhythmia characterized by CW wavefront rotation around the tricuspid annulus and critical dependence on the subeustachian isthmus (9,10,13). Atypical atrial Fl was defined as an atrial arrhythmia with a constant beat-to-beat intracardiac activation sequence different from either CCW or CW typical atrial Fl and not dependent on the subeustachian isthmus (13). The medial and lateral part of the subeustachian isthmus were defined as the zone between the inferior vena cava and the tricuspid annulus reaching from 4:00 (i.e., the coronary sinus os) to 6:00 and 6:00 to 7:00 o’clock on the LAO view, respectively. A multipolar body surface potential map was defined as the simultaneous occurrence of three or more extremes at a single time instant with an area of opposite polarity sandwiched between two extremes of equal polarity (18). Sudden extreme movement on a body surface potential map was defined as the rapid shift of an extreme over more than two lead sites on the chest (18). The electrically silent zone of the Fl wave cycle on the surface ECG was arbitrarily defined as the time interval in which both the highest positive and negative potentials were contained within a ±30 μV voltage window.
Data analysis and statistics
Endocardial data were analyzed by determining the activation time in each bipolar lead recording using on-screen digital calipers. The activation time in each bipolar signal was then related to the same time instant on the simultaneously recorded surface ECG, while the location of the endocardial recording site was visually translated from biplane fluoroscopic images to an adapted endocardial display of the right atrial endocardium which was previously published by Anderson and Becker (22).
Body surface mapping
Electrocardiographic data were analyzed qualitatively by inspecting the temporal changes in voltage distribution throughout the Fl wave cycle. This inspection entailed four steps. First, map pattern characteristics were evaluated by documenting the presence of qualitative features representative for complex underlying patterns of wavefront propagation (18): 1) multipolar map patterns; 2) pseudopod extensions, and 3) sudden extreme movement. Second, the spatial surface Fl wave morphology was compared instant-by-instant with the actual site of endocardial activation in the reentrant circuit. Third, the duration of the electrically silent zone in the Fl wave cycle was determined and associated with the underlying area of endocardial activation within the reentrant circuit. Fourth, distinct periods within the Fl wave cycle were identified at which stable voltage distributions occurred for a time interval of at least 20 ms. Subsequently, integral maps were computed over these selected time periods with stable voltage distributions. In addition, mean maps were calculated of all the CCW and CW typical Fl wave integral maps to document their distinct surface map morphology.
Quantitative data evaluation was carried out by subjecting the integral maps to a mathematical analysis using correlation coefficients in order to: 1) compare the intrapatient variability of the Fl wave pattern obtained at various periods of the Fl wave cycle with the purpose of mathematically determining the surface expression of circular wavefront propagation, and 2) assess the interpatient comparability of typical Fl wave patterns with the same direction of rotation using a jackknife procedure (18). The latter computations were performed using MATLAB (The MathWorks, Inc., Natick, Massachusetts) on a Sun Sparc Station 4 (Sun Microsystems, Inc., Palo Alto, California).
All data were expressed as mean ± SD. An unpaired two-tailed Student t test was performed if required with a p < 0.05 expressing statistical significance.
General features of atrial Fl
Counterclockwise typical atrial Fl was found in 13 patients, CW typical atrial Fl in 3 patients, both forms of typical atrial Fl in 4 patients and atypical atrial Fl in 1 patient (Table 1). The mean cycle length of CCW and CW typical atrial flutter was 228 ± 25 ms (range 200 to 300 ms) and 224 ± 23 ms (range 200 to 270 ms), respectively (Table 1). There was no significant difference in the cycle length of typical atrial Fl between patients with (229 ± 28 ms) or without structural heart disease (222 ± 16 ms) (p = 0.49). However, a significant difference in the cycle length of typical atrial Fl was observed in patients who were presently or recently treated with amiodarone (252 ± 31 ms) as opposed to patients not receiving any antiarrhythmic drug treatment (218 ± 14 ms) (p < 0.001).
Potential maps of typical atrial Fl
Map pattern characteristics
Visual evaluation of the instantaneous voltage distributions during typical atrial Fl showed a preponderance of dipolar patterns although multipolarity was observed at short periods within the Fl wave cycle in 53% and 57% of the CCW and CW Fl wave maps, respectively. These multipolar patterns were particularly evident when rapid surface map pattern transitions occurred, which coincided with the completion of right septal activation in both forms of typical Fl (60% of the multipolar Fl wave patterns). A striking incidence of pseudopod extensions (92%) and sudden extreme movement (100%) was additionally found with the two Fl rotations.
Comparison of surface morphology with endocardial activation-CCW rotation
Correlation of the temporal variation in surface map pattern and the endocardial activation sequence of CCW atrial Fl is shown in the following representative example. Figure 2 displays a series of nine potential maps obtained during CCW atrial Fl (panel C), which were each selected based on the local right atrial activation time recorded at sites 1 to 9 (panels A and B). The first 48 ms of the Fl wave cycle (negative component of Fl wave in lead II) are characterized by a stable pattern with a maximum and minimum on the right upper and middle lower anterior chest, respectively; this coincides with a caudocranial spread of activation in the right atrial septum (sites 1 to 3) and a proximal-to-distal activation sequence of the coronary sinus. Then a rapid spatial transition occurs (50 ms) to a pattern reflecting a complete reversal of the extreme locations (62 ms). These phenomena occur upon arrival of the activation wavefront at the anterior free wall of the right atrium near the superior crest (site 4). The multipolar map pattern observed at 50 ms most probably reflects both initial and terminal activation of the roof of the right (positive extreme at the middle anterior torso) and left atrium (positive potentials at the left upper anterior torso), respectively. Subsequently, when the impulse travels inferiorly down the trabeculated right atrium into the lateral part of the subeustachian isthmus (sites 5 to 9), the surface pattern remains stable (from 56 to 148 ms) (positive component in lead II) until the onset of the electrically silent zone of the terminal Fl wave at 150 ms. One may note that there is a subtle inferior shift of the minimum, resulting in more rightward inferiorly directed forces when activation of the lateral isthmus is nearly completed (site 9).
In the remaining 16 CCW Fl wave episodes, the spatial transition and subsequent complete reversal of the surface map pattern was observed during activation of the high right atrial septum or directly after completion of septal activation (i.e., from 2:00 to 12:00 o’clock on the LAO view). In addition, a proximal-to-distal activation of the coronary sinus coinciding with caudocranial septal activation was always found in these 16 CCW Fl wave episodes.
Comparison of surface morphology with endocardial activation-CW rotation
Figure 3 highlights a direct comparison of the characteristic sequential map pattern changes and the underlying site of endocardial activation within the macroreentrant circuit during CW atrial Fl. The early Fl wave distribution (first 76 ms) (positive component of Fl wave in lead II) shows a stable pattern with a maximum and minimum on the left middle and right upper chest, respectively. This pattern is the surface reflection of craniocaudal activation of the medial part of the anterior right atrial free wall (site 1) and the entire septum until the coronary sinus os (site 2). Then, after invasion of the Fl wavefront in the medial subeustachian isthmus, a first pattern transition takes place. The resulting map pattern features a slightly different location of the maximum on the middle anterior thorax while the minimum is now located on the middle left axilla (156 ms). This pattern remains stable for the next 80 ms until isthmus activation is completed (site 3). Note that: 1) activation of the coronary sinus occurs in a proximal-to-distal direction in concert with isthmus activation, 2) completion of right atrial septal activation coincides with the termination of the first part of the M-shaped positive component of the Fl wave in lead II, and 3) concurrent activation of the isthmus and floor of the left atrium is responsible for the second part of the M-shaped positive component of the Fl wave in lead II. Subsequently, the surface map morphology undergoes a second transition leading to a stable pattern with a maximum moving to the right upper chest and a slight shift of the minimum to the lower left axilla (from 170 to 220 ms). These surface phenomena result from caudocranial activation of the right atrial free wall (sites 4 to 9) (negative component of Fl wave in lead aVF).
All of the above described observations were also noted in the remaining 6 CW Fl episodes: 1) complete reversal of the voltage distribution occurred upon activation of the inferolateral right atrial wall directly after completion of isthmus activation (i.e., from 7:00 to 8:00 o’clock on the LAO view), 2) proximal-to-distal excitation of the coronary sinus in conjunction with medial-to-lateral isthmus activation, and 3) correspondence of the first and second part of the M-shaped positive component in the inferior leads of the standard ECG with activation of the right atrial septum (and roof of the left atrium) and combined activation of the isthmus and floor of the left atrium, respectively.
Electrically silent zone
There was a significant difference in the duration of the electrically silent zone of the typical Fl wave cycle when comparing CCW (55 ± 20 ms) with CW rotation (18 ± 8 ms) (p < 0.0001). The electrically silent zone in CCW and CW Fl episodes occurred during activation of the medial part of the isthmus (i.e., from 6:00 to 4:00 o’clock on the LAO view) and anterolateral free wall of the right atrium (i.e., from 11:00 to 12:00 o’clock on the LAO view), respectively.
Integral maps of typical atrial Fl
The Fl wave cycle in CCW typical Fl could be divided in two (10 of 17 episodes; 59%) or three successive time periods (7 of 17 episodes; 41%) with a stable surface map pattern of which integral maps were computed. A set of representative integral maps corresponding to a three-interval subdivision of the Fl wave cycle is shown in Figure 4. Complete pattern reversal can be observed when comparing the integral maps of the initial with the terminal part of the Fl wave (r = −0.93). It is interesting to observe that the downward deflection of the positive Fl wave component on the scalar ECG recorded at the E9 electrode position (left lower anterior chest) is characterized by a steep, followed by a gradual, decline in voltage. The first part of the gradual voltage decline corresponds to the time interval during which the third integral map was obtained and coincided with activation of the lateral subeustachian isthmus. A steep-to-gradual voltage decline of the downward deflection of the positive Fl wave component in the unipolar chest leads on the lower chest and the inferior leads of the standard ECG was noted in 11 of the 17 CCW Fl episodes (65%). This subset contained six of the seven CCW Fl episodes (86%) in which a three-interval subdivision could be obtained.
A division in two or three time periods with a stable map pattern was feasible in two (29%) and five of seven CW typical Fl episodes (71%), respectively. Figure 5 contains three representative integral maps of a CW Fl cycle, which were each based on a time interval with a stable voltage distribution. The integral maps corresponding with the initial and terminal part of the Fl wave show complete pattern reversal (r = −0.81) compliant with circular rotation of the underlying current source. The same CW Fl episode featuring the changes in the potential distribution over time was shown in panel A of Figure 3. It can be noted that the pattern of the initial, intermediate and terminal integral map show a high resemblance with the potential maps at 66, 156 and from 196 ms, respectively.
Mean maps of CCW and CW rotation
Mean integral maps of CCW and CW typical atrial Fl are demonstrated in Figure 6. The mean CCW integral map of the initial Fl wave is characterized by a maximum on the right upper anterior chest and a minimum on the lower left axilla (caudocranial activation of right atrial septum and proximal-to-distal activation of coronary sinus), the intermediate Fl wave by a maximum on the middle anterior chest and a minimum on the right upper thorax (craniocaudal activation of the right atrial free wall) and the mean CCW integral map of the terminal Fl wave by a maximum on the lower left anterior torso and a minimum on the right upper anterior chest (activation of the lateral subeustachian isthmus). In CW atrial Fl, the initial Fl wave produces a mean integral map featuring a maximum on the right middle front and a minimum on the right upper front (craniocaudal activation of the right atrial septum); the intermediate Fl wave generates a map with a maximum on the right middle chest and a minimum on the middle left axilla (activation of subeustachian isthmus and proximal-to-distal activation of coronary sinus), and the terminal Fl wave leads to a mean integral map showing a maximum on the right upper front and a minimum on the left middle front of the torso (caudocranial activation of right atrial free wall). There appeared to be a high spatial resemblance of the mean integral maps corresponding to the intermediate CCW Fl wave and the initial CW Fl wave. Similarly, the integral maps of the initial CCW Fl wave and the terminal CW Fl wave are morphologically very compatible.
Mathematical comparison of the initial and terminal integral map performed for each of the 17 CCW typical Fl episodes shows clear evidence of the spatial map pattern reversal resulting from the underlying rotation of the endocardial impulse and is reflected by a high negative correlation (mean r = −0.87 ± 0.08; range −0.70 to −0.96). Similar findings were obtained for the seven CW typical Fl episodes although the negative correlations were somewhat lower (mean r = −0.55 ± 0.29) and more variable (range −0.05 to −0.95) compared with the results obtained with CCW Fl.
Comparison among either CCW or CW typical Fl episodes obtained in different patients demonstrates a high degree of interpatient pattern uniformity irrespective of the presence or absence of structural heart disease. With CCW Fl, correlation coefficients of initial integral maps ranged from 0.71 to 0.97 (mean r = 0.89 ± 0.09), intermediate integral maps from 0.67 to 0.97 (mean r = 0.86 ± 0.10) and terminal integral maps from 0.72 to 0.94 (mean r = 0.83 ± 0.09). Results with CW flutter yielded r values for initial integral maps ranging from 0.64 to 0.90 (mean r = 0.80 ± 0.09), for intermediate integral maps ranging from 0.67 to 0.97 (mean r = 0.82 ± 0.12) and for terminal integral maps ranging from 0.71 to 0.81 (mean r = 0.77 ± 0.05).
Atypical atrial Fl
The one episode of atypical Fl showed three successive periods in the Fl wave cycle during which stable and predominantly dipolar map patterns were observed. The corresponding integral maps are shown in Figure 7. The initial 42 ms demonstrate a maximum and minimum on the middle anterior chest and upper left back, respectively (left panel). This time period is associated with: 1) craniocaudal right septal activation followed by activation of the coronary sinus os, and 2) a predominantly positive Fl wave polarity in the inferior leads of the standard 12-lead ECG. Then, a first transition of the surface map pattern takes place and results in a slightly more superior position of the maximum on the anterior chest and a minimum located on the lower back (middle panel). This surface pattern remains present for 52 ms and coincides with a complex endocardial activation sequence featuring concurrent craniocaudal activation of the entire right atrial free wall, activation of the subeustachian isthmus and proximal-to-distal activation of the coronary sinus. Then, a second surface map pattern transition causes the maximum and minimum to move to the left upper and middle anterior thorax, respectively (right panel). This pattern remains stable for 38 ms until the electrically silent zone occurs at 132 ms. No endocardial activity was noted in the right atrium or coronary sinus during this last phase of the Fl wave cycle. Thus, the terminal map pattern must be caused by left atrial activation, most probably by caudocranial wavefront propagation along the left atrial free wall. In addition, we speculate that the electrically silent zone (132 to 170 ms) on the surface ECG is associated with activation of the roof of the left atrium. This atypical Fl with a probable underlying left atrial macroreentrant circuit also showed potential maps featuring sudden extreme movement, pseudopod extensions and multipolarity during rapid pattern transitions. Furthermore, complete spatial reversal of the atypical Fl wave map pattern can be noted when comparing the initial with the terminal integral map (r = −0.77).
This report contains the first comprehensive evaluation of the multiple lead surface ECG morphology of typical atrial Fl and was aimed at providing new insight into the genesis of the Fl wave and outlining the potential clinical utility of surface mapping techniques to characterize type and direction of Fl wave rotation. By performing simultaneous body surface and endocardial mapping during 17 episodes of CCW typical atrial Fl, seven episodes of CW typical atrial Fl and one episode of atypical atrial Fl in 20 patients selected for ablative therapy, we found that the temporal changes in Fl wave morphology on the body surface can be directly related to the successive activation of major components in the atrial macroreentrant circuit. The CCW and CW typical Fl wave cycle can each be divided in up to three time periods at which stable surface map patterns are observed; with CCW Fl, these periods correspond to caudocranial activation of the right atrial septum in conjunction with proximal-to-distal activation of the coronary sinus (initial Fl wave), craniocaudal activation of the right atrial free wall (intermediate Fl wave) and activation of the lateral subeustachian isthmus (terminal Fl wave) (Fig. 2); with CW flutter, the time periods can be attributed to craniocaudal excitation of the right atrial septum (initial Fl wave), excitation of the isthmus together with proximal-to-distal excitation of the coronary sinus (intermediate Fl wave) and caudocranial excitation of the right atrial free wall (terminal Fl wave) (Fig. 3). Potential as well as integral maps of the initial and late period of either the CCW or CW typical Fl wave cycle showed complete spatial pattern reversal due to impulse propagation in opposite directions with excitation of the septal versus the free wall section of the right atrial Fl circuit (Fig. 2 to 5). A high level of quantitative interpatient concordance in both CCW and CW typical Fl wave body surface integral map patterns was found despite the presence of a variety of organic heart conditions in two thirds of the patient population. The atypical Fl episode was believed to be associated with a left atrial reentrant circuit and demonstrated clearly different integral maps of the initial, intermediate and terminal period of the Fl wave cycle as compared with the integral maps acquired during CCW or CW typical atrial Fl (Fig. 6 and 7).
The scalar surface ECG of typical atrial Fl and its relation to the intraatrial activation sequence
Puech et al. (5) showed that in patients with CCW typical atrial Fl, the positive component of the Fl wave in the inferior leads of the 12-lead ECG represents craniocaudal activation of the right atrial free wall, while the negative component of the Fl wave relates to caudocranial propagation of the impulse along the right atrial septum. Recently, multisite endocardial mapping of typical atrial Fl with both directions of rotation corroborated the earlier findings regarding the nature of the scalar CCW Fl wave (11,13). These studies also showed that a reversed phenomenon occurred with CW Fl, i.e., craniocaudal septal and caudocranial lateral right atrial activation while the positive and negative Fl wave components in the inferior leads were inscribed. Saoudi et al. (11) demonstrated that coronary sinus activation in both types of Fl rotation occurs in a proximal-to-distal direction and that the notch at the beginning of the plateau phase of the CW Fl wave in the inferior leads (i.e., the second part of the M-shaped component) is caused by excitation of the coronary sinus (or floor of the left atrium). We have obtained similar findings and noted that the first part of the M-shaped component correlates with the craniocaudal activation of the right atrial septum (and roof of the left atrium). We additionally found that CCW impulse propagation in the lateral isthmus was often associated with distinct scalar features in the lower chest leads and the inferior leads of the standard ECG. In almost two thirds of the CCW Fl episodes, the positive Fl wave component demonstrated a steep, followed by a gradual, voltage decline with the initial part of the gradual voltage decline occurring simultaneously with activation of the lateral isthmus (Fig. 4).
Role of the bystander atrium in generating the electrocardiographic Fl wave
Sequential epicardial mapping of CCW and CW macroreentrant activity in the canine sterile pericarditis model has demonstrated that the morphogenesis of the surface Fl wave is predominantly determined by the activation sequence of the left atrium, which appeared to act electrically as a passive bystander (23). A similar explanation for the dominant Fl wave polarity in experimental Fl due to right atrial crush injury (24) as well as the typical form of clinical atrial Fl (5,11,13) has been proposed by others. Schoels et al. (25) stated that the Fl wave polarity in experimental atrial Fl is related to the direction of activation of the bulk of atrial tissue. Assuming that the terms polarity or dominant component of the Fl wave relate to a deflection of the waveform with a relatively high amplitude and rapid increase or decrease in voltage, we feel that both the direction of activation in the bystander atrium and the synchronization of right and left atrial activity (i.e., the amount of simultaneously activated right and left atrial tissue mass) determine the polarity or dominant component of the scalar Fl wave. Thus, synchronized excitation of the right atrial septum and left atrium in either a caudocranial or craniocaudal direction will be responsible for the Fl wave polarity in the typical form of atrial Fl where the reentrant circuit is located in the right atrium. Although the remaining “nondominant” part of the scalar Fl wave in typical atrial Fl, generated by sequential activation of the right atrial free wall in a craniocaudal or caudocranial direction, may be difficult to evaluate visually due to its relatively low amplitude, it should be emphasized that the use of spatial ECG sampling with a high signal-to-noise ratio does enable a detailed instant-by-instant representation of the “nondominant” Fl wave voltage distribution on the human torso. Indeed, the ability to spatially capture nearly the entire Fl wave distribution using the surface mapping approach additionally unveiled a striking pattern correspondence of the mean integral maps belonging to the intermediate CCW Fl wave and the initial CW Fl wave as well as the mean integral maps belonging to the initial CCW Fl wave and the terminal CW Fl wave (Fig. 6). This intriguing finding reiterates the importance of the direction of wavefront propagation in either the right atrial free wall or septum (and left atrium) in determining the spatial distribution of typical atrial Fl waves on the torso surface.
Biatrial wavefront propagation and cancellation of electromotive forces
Boineau et al. (26) obtained canine epicardial mapping data during naturally occurring CCW and CW right atrial Fl. They showed that synchrony in direction of wavefront propagation and cancellation of opposing wavefront propagation determine the presence or near absence of electromotive forces on the surface ECG, respectively. During clinical observations using multisite right atrial endocardial mapping in patients with atrial fibrillation, Roithinger et al. (27) noted that the amplitude and polarity in the V1 lead tracing were distinctly related to the degree and direction of organized activation of the in close proximity situated trabeculated right atrium, despite the cancellation effects created by simultaneously occurring disorganized activation on the right atrial septum. In the intermediate period of the CW typical Fl cycle, while concomitant activation of the subeustachian isthmus and concentric activation of the coronary sinus (and floor of the left atrium) is taking place in opposite directions, we were able to obtain a reproducible dipolar map pattern on the torso surface (Fig. 3, 5 and 6). Thus, two simultaneous biatrial wavefronts travelling in opposite directions do not cause complete cancellation but rather an interaction between their respective electromotive forces that result in a distinct dipolar surface map distribution. Although one may even expect to see multipolarity of the surface map pattern as a reflection of the aforementioned two activation wavefronts, insufficient separation of the electrical sources most probably precluded the observation of such phenomena.
The electrically silent zone of the Fl wave cycle on the surface ECG
In the Langendorff-perfused canine heart, Allessie and coworkers (28) performed endocardial mapping of pacing and acetylcholine induced left and right atrial Fl. These authors observed that continuous electrical activity was only present in one of the two atria (i.e., the atrium containing the Fl circuit) and that the amount of instantaneously activated myocardium was less than 5% during half of the Fl wave cycle. They suggested that the endocardially quiescent part of the Fl wave cycle was due to activation through a narrow isthmus of atrial myocardium that was part of the macroreentrant circuit and responsible for the “isoelectric” segment on the surface ECG. Subsequent clinical observations in patients with CCW typical atrial Fl corroborated these experimental results in that the “isoelectric” segment on the surface ECG relates to activation of the low right atrium (i.e., the subeustachian isthmus) (7,13). Similarly, we found with CCW typical Fl that activation of the small muscle mass of the medial part of the subeustachian isthmus generates a low amount of electromotive forces, which results in a period of relative electrical silence on the surface ECG (i.e., voltages within a ±30 μV window). In addition, it appeared that the timing of the silent zone on the surface ECG is not only related to the absolute muscle mass of a given segment of the right atrial Fl circuit but also to the time at which passive activation of the bystander left atrium occurs. In fact, the electrically silent zone in both rotations of typical atrial Fl can be observed just before the onset of passive left atrial activation. While the silent zone in CCW typical Fl is noted to occur during activation of the medial isthmus when caudocranial activation of the bystander left atrium is about to commence via a lower transseptal route, the silent zone in CW typical Fl appears to coincide with terminal activation of the anterolateral free wall of the right atrium at the level of the superior crest just before the start of craniocaudal excitation of the left atrium via Bachmann’s bundle. The fact that the duration of the electrically silent zone was significantly longer in CCW (55 ± 20 ms) as opposed to CW typical Fl (18 ± 8 ms) most probably relates to differences in mass and possibly conduction velocity due to the anatomic architecture of the rim of atrial muscle at the level of the isthmus and anterolateral free wall of the right atrium.
Because typical atrial Fl is characterized by continuous macroreentrant activity around the tricuspid annulus, per definition the Fl waveform on the surface ECG will not contain a true isoelectric interval. There are, however, distinct time instants in both the CCW and CW typical Fl wave cycle when a minimal amount of electromotive forces is generated at which time the surface ECG can be considered to be electrically silent. In our approach for baseline correction and definition of the zero voltage level, we decided to select such a time instant directly preceding the negative inscription of the Fl wave in accordance with the methods described by Kawano et al. (19). Although unlikely, due to the high quantitative interpatient concordance in both CCW and CW Fl map patterns, we cannot exclude that minor artefacts may have been introduced by the visually guided selection of the zero voltage.
Due to the lower clinical prevalence of CW as compared with CCW typical atrial Fl (11–13,17), we were only able to include a restricted number of CW Fl episodes in the analysis of this report. Therefore, one may question whether the current findings on CW typical Fl can be extrapolated to larger patient populations. However, we feel that the reproducibility of the visual CW Fl potential map analysis on a patient-to-patient basis, the high quantitative interpatient CW Fl map pattern correspondence and a comparable distribution of patients with and without associated heart disease in the CW typical Fl group (Table 1) justify the use of the currently developed set of mean CW typical Fl integral maps as a reference frame to guide future analysis.
The use of surface maps to improve detection of typical atrial Fl or optimize electrocardiographic characterization of Fl circuit location may be particularly useful in the clinical scenario where typical atrial Fl may not be recognized and considered for ablative therapy because the 12-lead ECG pattern during CW rotation is not “close to classic” (11) or when the 12-lead ECG pattern of a truly atypical atrial Fl may be mistakenly labeled as a typical Fl that would be amenable for ablation (13). It was beyond the scope of this study to compare the performance of 62 versus 12-lead ECG recordings in their ability to detect and classify typical atrial Fl. Nevertheless, future prospective studies designed to address this issue are warranted in order to fully determine the clinical role of spatial 62-lead ECG sampling in the management of atrial Fl.
Clinical relevance and future directions
The present data show that application of 62-lead ECG mapping to characterize typical atrial Fl offers a comprehensive spatial representation of the specific CCW and CW typical Fl wave morphology on the human torso surface. This spatial ECG sampling approach allows for direct correlations of the instantaneous surface map pattern during a given time period in the Fl wave cycle with activation of distinct components in the right atrial macroreentrant circuit. In addition, the mean integral maps of the three main stages in the CCW or CW typical Fl wave cycle present a unique reference set of map patterns that can be used for more sophisticated and less ambiguous clinical classification of the Fl type and its direction of rotation than currently feasible with the existing algorithms based on scalar features of the 12-lead ECG. Clinical application of the reference set of mean CCW and CW typical Fl wave integral maps may also lead to improved noninvasive methods of patient selection so that treatment strategies can be better tailored regarding the optimal choice of therapy before invasive electrophysiologic studies are carried out.
At present, we are unaware about the specificity of surface mapping techniques in being able to discriminate typical from atypical atrial Fl with either a nonisthmus dependent rightsided circuit (13) or a circuit located in the left atrium (29,30). Although we did find clearly different Fl wave integral map patterns in the one example of atypical Fl believed to be associated with a left atrial circuit, on a larger scale we do not know whether the spatial map pattern of atypical Fl will contain particular features specific to a given and potential leftsided circuit location. However, given the invasive nature and risk of transseptal catheter manipulation when targeting ablation of anticipated leftsided circuits, it is obvious that the development of noninvasive techniques that allow predictions about atypical Fl circuit locations may be of critical importance not only to guide the process of clinical decision making but also to accelerate and focus catheter activation and entrainment mapping to a confined area of interest. These issues should be the subject of study in future clinical investigations so that improved and more complete multiple-lead ECG classification algorithms become available that cover the entire clinical spectrum of atrial Fl.
We thank Mark Potse, MSc, and Andre C. Linnenbank, PhD, for their assistance in data processing.
☆ This study was, in part, supported by the Royal Netherlands Academy of Arts and Sciences and the National Institutes of Health (HL09602 and RO1-HL55227).
- left anterior oblique
- right anterior oblique
- Received August 4, 1999.
- Revision received November 9, 1999.
- Accepted December 29, 1999.
- American College of Cardiology
- ↵Jolly WA, Ritchie WT. Auricular flutter and fibrillation. Heart 1910–1911;2:177–221.
- Lewis T,
- Drury A.N,
- Iliescu C.C
- ↵Lewis T, Feil HS, Stroud WD. Observations upon flutter and fibrillation. Part II. The nature of auricular flutter. Heart 1918–1920;7:191–245.
- Puech P,
- Latour H,
- Grolleau R
- Olshansky B,
- Okumura K,
- Hess P.G,
- Waldo A.L
- Feld G.K,
- Fleck P.R,
- Chen P.S,
- et al.
- Lesh M.D,
- Van Hare G.F,
- Epstein L.M,
- et al.
- Olgin J.E,
- Kalman J.M,
- Fitzpatrick A.P,
- Lesh M.D
- Cosio F.G,
- Arribas F,
- Lopez-Gil M,
- Nunez A
- Waldo A.L,
- MacLean W.A.H,
- Karp R.B,
- et al.
- Kalman J.M,
- Olgin J.E,
- Saxon L.A,
- et al.
- SippensGroenewegen A,
- Peeters H.A.P,
- Jessurun E.R,
- et al.
- Linnenbank AC. On-Site Recording, Analysis and Presentation of Multichannel ECG Data. PhD Thesis, University of Amsterdam, the Netherlands, 1996.
- Anderson R.H,
- Becker A.E
- Okumura K,
- Plumb V.J,
- Page P.L,
- Waldo A.L
- Feld G.K,
- Shahandeh-Rad F
- Schoels W,
- Offner B,
- Brachmann J,
- et al.
- Allessie M.A,
- Lammers W.J.E.P,
- Bonke I.M,
- Hollen J
- Jais P,
- Haissaguerre M,
- Shah D.P,
- et al.