Three-dimensional Mapping of the Common Atrial Flutter Circuit in the Right Atrium
Background The full circuit of common atrial flutter using conventional methods of sequential or multielectrode activation mapping is not completely understood.
Methods and Results We performed three-dimensional right atrial endocardial activation mapping during common counterclockwise atrial flutter in 17 patients (16 men, 1 woman; mean age, 53±11 years) by using the Cordis-Biosense EP Navigation system and assessed the distribution of estimated conduction velocities and double and fractionated potentials. ECG flutter wave morphologies were compared with activation patterns. Points (91±29) were sequentially acquired covering 88±11% of the flutter cycle length of 239±22 ms. A wide and variable posterior zone of double and fractionated potentials coincided with blocking and colliding wave fronts and formed the posterior limit of the circuit. A progressively widening septal (sep) wave front ascending from just beyond the coronary sinus ostium, passed cranially as a broad front anterior to the superior vena cava (SVC) in 14 patients, whereas fusion around the SVC formed the superior (sup) limb of the circuit in 3. Bounded anteriorly by the tricuspid valve, the wave front descended down the lateral (lat) aspect of the right atrium before completing the circuit in all cases through the inferior vena cava–tricuspid annulus isthmus. The estimated conduction velocity in the medial isthmus (0.6±0.3 m/s) was lower than in the other limbs of the circuit (sup=1±0.5 m/s, lat=1±0.5 m/s, sep=0.9±0.4 m/s, P=.05). Double and fractionated potentials were constant and more prevalent in the posterior right atrium. ECG flutter wave morphology did not correlate with three-dimensional activation maps.
Conclusions Interindividual variations occur in the right atrial circuit of common atrial flutter, with constant activation through the cavotricuspid isthmus. A variable zone of block forms the posterior limit. Fusion around the SVC can occur, and ascending medial septal activation does not follow a consistent pattern.
Activation mapping has been a mainstay for understanding many cardiac arrhythmias, particularly supraventricular and reentrant tachycardias.1 The limitations of conventional techniques involving single bipolar/unipolar sequential mapping or multielectrode/multicatheter mapping based on fluoroscopic correlations are, however, particularly obvious when attempting to relate to a three dimensional structure.
Previous studies2–8 have shown that typical atrial flutter is a stable macroreentrant tachyarrhythmia with a counterclockwise circuit revolving around anatomic and functional blocks in the right atrium. A common isthmus between the inferior vena cava and the tricuspid annulus critical to this arrhythmia has been identified by ablation techniques.6,9–12 Both conventional activation as well as entrainment mapping have been used to infer the participation of parts of the atria,13 but the full circuit remains relatively ill defined. In this study we describe the three-dimensional (3D) activation pattern in the right atrium (RA) during typical atrial flutter by using a new technique of correlative 3D mapping with particular emphasis on superior and septal-medial right atrial activation.
Seventeen patients referred for radiofrequency (RF) catheter ablation of symptomatic typical atrial flutter were included in this study after informed consent was obtained. The study was approved by our institutional review board. All antiarrhythmic drugs were stopped at least 4 half-lives before the study except for amiodarone in 8. Sixteen of these patients were men; their mean age was 53±11 years (range, 41 to 74). Two patients had structural heart disease, including 1 with dilated cardiomyopathy and 1 with mitral regurgitation. All patients were required to have a typical surface ECG flutter wave morphology, which included sawtooth flutter waves negative in leads II, III, and aVF. Patients with nonsustained atrial flutter, surface ECGs not suggestive of typical atrial flutter, or exhibiting frequent degeneration into atrial fibrillation were excluded. None of the patients had undergone previous catheter ablation or cardiac surgery.
Introducer sheaths (8F and 9F) were placed in the right (and if required, left) femoral veins to allow introduction of mapping and reference catheters. The examination was performed on fasting patients. Light sedation was administered with intravenous midazolam and/or buprenorphine as necessary. Two patients who were in sinus rhythm before the procedure underwent rapid burst pacing from the proximal coronary sinus or low right atrium to induce typical atrial flutter. The preliminary part of the study protocol required confirmation of caudocranial septal and craniocaudal lateral right atrial activation with a duodecapolar Halo catheter (Cordis-Webster) in 16- and a 14-pole custom-designed woven dacron catheter (Bard Inc) in one patient.
The EP Navigation System includes a location pad, the Carto processor, a monitor, and workstation (Silicon Graphics Inc) as well as sensor-equipped mapping catheters (Cordis EP Navigation catheters). The location pad with its three pods mounted on a triangular frame is attached to the underside of the fluoroscopy table and emits a weak electromagnetic field allowing a passive sensor incorporated in the catheter tip of an 8F EP Navigation catheter to sense and transmit these signals back to the processor unit. A triangulation algorithm is used to precisely determine the position of the catheter tip in this electromagnetically coded space. One such catheter is placed in a suitable reference position, in this protocol within the proximal coronary sinus so as to record a large and stable atrial electrogram and a minimal ventricular electrogram. A second such catheter is used as a roving map catheter in the right atrium. The real-time spatial coordinates of the reference catheter (which, like the map catheter, vary phasically both with respiration and cardiac motion) are used to provide a dynamic corrective reference for the spatial location of the map catheter tip. This allows correct location and relocation of the position of the map catheter tip despite absolute intrathoracic positional changes (due to respiration and cardiac motion). The tip of the map catheter is depicted as a 3D icon on the screen of the monitor, and its position and orientation in space can be ascertained both by changes in the orientation or position of the icon as well as by manipulating the background 3D grid (with the workstation mouse) to alter the viewing perspective. With a rapid update rate, such a capability can allow a certain amount of nonfluoroscopic manipulation inside the reconstructed cardiac chamber.
A 3D map of the right atrium during atrial flutter is constructed by sequential mapping, ie, incorporating electrogram-derived activation times as well as spatial coordinates from each of multiple sites over the right atrial endocardium. At a given location, 2 to 3 seconds of electrograms are retained in memory and if variations in successive cycles of local activation times as well as spatial coordinates are within specified limits indicating both catheter and rhythm stability, this data point is incorporated in the map. The software uses a STAR algorithm to construct a polyhedral 3D structure using triangular panels from the spatial coordinates obtained at different locations. This algorithm incorporates a programmable distance threshold for interpolation that was progressively decremented during data acquisition to achieve a greater uniformity of point distribution for this study. The panels are color coded by assigning the colors of the rainbow to relative local activation timings (relative to a reference electrogram fiducial point). Activation times as well as the reference electrogram fiducial point are automatically detected according to programmable criteria. The accuracy of this system has been tested both in vitro and in vivo and found to be reproducible and accurate.14
For the first two patients the peak of the surface ECG (lead II) R wave was used as the reference fiducial point after verification of stable and fixed 2:1 atrioventricular conduction. The minimum electrogram slope (ie, maximum negative dv/dt) of the proximal coronary sinus atrial electrogram was selected in the next 15 patients. Bipolar (in 15) and unipolar tip electrograms (in the initial 2 cases) were acquired for mapping purposes with bandpass filter settings of 30 to 400 Hz and 0 to 400 Hz, respectively. In some patients both unipolar and bipolar signals were simultaneously stored in the memory with unipolar signals used to edit/assign local activation times in the presence of double potentials. A double potential was defined as a double spike (bipolar) or double deflection (unipolar) electrogram with an interval of at least 20 ms measured between the points with the minimum dv/dt in each spike or deflection. This interval for each double potential complex was measured in milliseconds with an on screen caliper. Bipolar electrograms with a duration longer than 50 ms and consisting of continuous fractionated electrical activity or with multiple (>3) deflections were also identified.4 The positions of points recording stable, consistent (≥3 consecutive cycles) double potentials or fractionated potentials fulfilling the above definition were marked graphically on the 3D maps by prominent separately colored roundels. The distribution of double potentials and fractionated potentials was assessed during atrial flutter in the context of their role as markers of local conduction block or delay. Analysis of acquired data therefore included the following steps:
(1) All the acquired map data points were manually scrutinized for fulfillment of local activation time and location stability criteria. Screening requirements were a variation <5 ms and a tip location variability of <3 mm on successive beats. Points not fulfilling these criteria were excluded from the database and the reconstructed 3D map.
(2) Reference electrogram fiducial points were manually checked; if incorrectly annotated these data points were excluded from the map database.
(3) Map electrogram local activation times were also verified manually and corrected if necessary.
(4) Double potentials and fractionated potentials were marked by differently colored roundels on the 3D map. They were assigned to the following areas of the RA: (1) posterior RA, (2) superior vena cava (SVC), (3) pericoronary sinus ostium region, (4) septal RA, (5) inferior anterolateral RA, (6) superior anterolateral RA and (Fig 1⇓). These areas were determined on the reconstructed 3D map by correlation with fluoroscopic landmarks, electrogram markers (eg, the His bundle electrogram, A-V electrograms at the tricuspid valve annulus) and the reconstructed map itself (see later). Two points closer than approximately 4 mm in any dimension were considered as one for quantitative analysis.
(5) The volume of the reconstructed 3D map was automatically calculated.
(6) Software-generated animated propagation maps were superimposed on the reconstructed anatomic contours. The animation maps were qualitatively analyzed for macro activation patterns as well as colliding activation wave fronts, particularly in relation to anatomic landmarks/obstacles such as the orifices of the vena cava and the tricuspid annulus. Activation fronts were considered to be within the reentrant circuit when they were part of the spatially shortest route encompassing the full range of mapped activation times and returned adjacent to or within close proximity of the site with the earliest activation. Wave fronts colliding with each other (either traveling in opposite directions or with an angle between them of more than 90° in two dimensions) with significantly disparate activation times (>20 ms) were considered presumptive of local block at the site of collision. Each limb of divergent wave fronts was evaluated contextually and on the basis of its destination. Wave fronts converging at an angle of less than 90° and without significantly disparate activation times were considered to fuse. Activation times were also compared anterior and posterior to the SVC on a line perpendicular to the advancing wave front using the means of activation times of three points relatively equally spaced around this line.
Activation fronts as well as double and fractionated potentials were assigned to specific anatomic locii on the basis of the reconstructed 3D map of the RA. In all cases the orifice of the tricuspid valve and the SVC were easily recognizable−the former by an appropriately placed defect in the map contour and the latter by a crownlike extension (representing electrical activity in this proximal great vessel segment) of the cranial aspect. Similarly the inferior, posterior, and leftward extremity of the reconstruction represented the proximal coronary sinus and its ostium (Fig 1⇑). Because we used a proximal coronary sinus atrial electrogram as a time reference, our mapped and demarcated activation cycle commenced at the coronary os and terminated in or around the coronary ostium–tricuspid isthmus, thus locating the “frameshift” region of the map there and precluding detailed analysis of activation in this region. We could thus demarcate posterior as well as septal-medial regions of the RA beyond the ostium of the coronary sinus. The septum was further subdivided into roughly equal vertical thirds: anterior (ventral), central, and posterior (dorsal), as judged from a superior (cranial) perspective for analyzing the sequence of activation in this area.
(7) An estimate of regional conduction velocity was calculated from the ratio of distances between points located in the mainstream of well-defined activation wave fronts and differences in activation times. Thus four points forming a quadrilateral within such a wave front were selected. The difference of the means of the activation times of the two points forming each side of the quadrilateral parallel to the advancing flutter wave front was divided by the estimated distance between the midpoints of these sides (23±8 mm). This distance varied because of the method of data acquisition utilized but points too close together or too far apart were excluded. Points were also selected to minimize marked intervening curvatures. For double potentials, the activation time of the potential corresponding to the orthodromically proximal wave front was used; in case of fractionated potentials the maximum negative dv/dt activation time was used for estimating conduction velocity. Conduction velocities were assessed in this manner in the medial part of IVC–tricuspid isthmus, the mid septum, the superior RA (anterior to the SVC), and the lateral RA. Because of the heterogeneous and complex activation observed in the posterior RA, conduction velocity could not be estimated.
Relation to Surface ECG Morphology
Interindividual variations in the surface ECG morphology were compared with activation mapping data. For 16 of 17 patients, 12-lead recordings with 3:1 or higher AV ratios were available. Two independent observers, blinded to the results of the animated propagation maps compared the following ECG features in lead III: (1) amplitude of the summed negative and positive components of the F wave, (2) the slopes of the flutter wave sharp component, and (3) of the plateau phase as well as (4) the ratio of the duration of the plateau (Tpl) to the duration of the F wave (Tf) (Fig 2⇓). These were then related to the results of 3D mapping.
All patients subsequently underwent catheter ablation of the inferior vena cava–tricuspid annulus (IVC-TA) isthmus, with the use of methods described in earlier reports.6,12 During ablation, isthmus electrograms were in all cases confirmed to be centered on the surface F-wave plateau. Standard quadripolar 4-mm, deflectable-tip electrode catheters were used in conjunction with a Cordis-Stockert RF generator. Closed loop temperature controlled and sequential point by point RF applications were performed in a linear fashion.
Each continuous variable was expressed as a mean±SD, and comparisons were made using the Kruskall-Wallis nonparametric test, the Fisher’s exact test, the W-paired Wilcoxon test, Student’s t test ,and ANOVA as considered appropriate. A value of P<.05 was considered significant.
Three-dimensional Activation During Common Atrial Flutter
Three-dimensional maps of the RA during typical atrial flutter were successfully generated for all 17 patients, and a mean of 91±29 points were acquired per map covering 88±11% of the flutter cycle length in terms of the range of local activation times mapped. The mean reconstructed RA volume was 77±36 mL (range, 36 to 188) and the mean atrial flutter cycle length was 239±22 ms (range, 200 to 268). There were no significant differences in cycle lengths (234±28 ms on and 243±15 ms off; P=NS) and RA volume (62±26 mL on versus 91±40 mL off; P=NS) between patients on and off amiodarone. All patients underwent successful RF catheter ablation after completion of the mapping procedure with a mean of 10±7 pulses (range, 2 to 29; median, 8). There were no significant side effects.
In all cases, counterclockwise activation of the right atrium was limited anteriorly by the boundary of the tricuspid valve. However, there was no similar discrete posterior boundary (Fig 3⇓). A posterior area of double potentials was noted in each map consistent with the presumed anatomic location of the crista terminalis, forming the posterior limit of the circuit as previously described.13 However, this posterior area was ragged and covered a zone that reached a maximum width of 1.7±0.8 cm (Fig 4⇓). Cranial activation during flutter consisted of a broad limb extending uniformly between the SVC and the superior tricuspid annulus in 14 cases before continuing caudocranially down the lateral RA wall (Fig 5a⇓). In three patients, the activation wave front was noted to spread both anteriorly and posteriorly to the SVC then fuse around the SVC before activating the lateral RA craniocaudally. In these patients there were gaps in the distribution of double potentials in the posterior RA behind the SVC (as also in other patients) and the region posterior to the SVC was activated just a little earlier (5 to 25 ms) than the anterior (Fig 5b⇓). The flutter cycle lengths were not different from the rest (233±25 versus 240±22 ms, P=NS). There were no significant differences in conduction velocities, and only one of them was receiving amiodarone.
In the septum there was caudocranial activation in all cases, without a constant pattern: In 6 cases a symmetrical wave front swept upwards (Fig 6⇓), in 4 the central septum was activated early and followed by centrifugal spread; in another 4 the anterior (ventral) septum was activated earlier with subsequent spread; whereas in 1 patient the posterior (dorsal) septum was activated before the anterior septum. There was random and patchy activation in 2.
In all cases, however, the flutter circuit was completed by activation through the IVC-TA isthmus and correspondingly activation was noted to occur to varying extents in the opposite direction posterior to the IVC before blocking. On the animated propagation map, one or more major activation wave front collisions with disparate activation times were observed in all cases in the posterior RA coinciding in some cases very well with locations where double potentials were recorded. The posterior RA region was consistently noted to be activated by diverging relatively horizontal and asynchronous activation wave fronts in contrast to activation in the vertical plane in the lateral and septal regions as previously described.
Regional Conduction Velocity Estimates
Table 1⇓ summarizes the data. Conduction velocity was significantly lower in the medial IVC-TA isthmus (0.6±0.3 m/s) when compared with that in the other regions. While this was slowest in 13/17 (76%) it was not so in 4 cases (2 of these 4 patients−50%−and 6 of the previous 13% to 46%−were receiving amiodarone). Though the velocities were lower in patients who were still under the effect of oral amiodarone, this was not statistically significant.
Regional Distribution of Double Potentials and Fractionated Potentials During Atrial Flutter
Double potentials and fractionated potentials were found at 22±8% and at 16±11% of points, respectively, in the RA (Figs 3⇑ and 4⇑); their incidence, extent, and characteristics are summarized in Table 2⇓. In some instances, “double potentials” with a Wenckebach-type relationship or frank dissociation were noted in the proximal SVC.24 A general pattern of posterior right atrial as well as septal and pericoronary ostial double potential clustering was noted (Fig 3⇑). The incidence, extent, and interspike DP intervals in the posterior RA were significantly greater than in the lateral RA and the anterior RA regions. There were also no significant differences between patients receiving and those not receiving amiodarone.
Comparison of Surface ECG Flutter Waves With Three-dimensional RA Maps
Morphologies of individual patients’ flutter waves varied (Table 3⇓), but there was no evident correlation between the RA maps of patients and the surface ECG variables.
Three-dimensional Mapping of Common Atrial Flutter
Recent anatomic studies and those derived from activation and entrainment data have improved our understanding of the circuit of common atrial flutter, particularly in the low septal region.6,7,15-24 The present study using a new technique of reconstructive 3D mapping of the right atrium in humans completes previous data by providing panoramic and comprehensive (360°) mapping during common type I right atrial flutter.
A mean of 91 points were used to reconstruct the RA and delineate the 3D activation patterns of the right atrium during flutter, which is similar to the mapping density in most experimental studies. Anteriorly RA activation is obviously limited by the tricuspid valve annulus. However unlike some previous descriptions in animal models,17 the activation front covers not just a peritricuspid band but in fact a wide area bordered by the annulus. This may explain why attempts in such models to terminate the flutter by limited ligatures in the lateral right atrial region were unsuccessful.17 In contrast, the posterior limit is more uneven as inferred from the constantly present but ragged distribution of double potentials in this area. Such double potentials have been demonstrated to be indicative of conduction block by entrainment techniques.13,18 19,23 Moreover, analysis of animated activation maps indicates diverging asynchronously traversing oblique and horizontal wave fronts in the posterior RA, suggesting a nonuniform and nonlinear zone of block as the posterior limit. This area corresponds broadly to the expected position of the crista terminalis,13 although the irregular outline and scattered distribution suggest a functional area of block beyond this discrete anatomic structure.
The precise delineation of the cranial limb of the flutter circuit has not been previously reported. This study shows that activation in this region consists of a broad front extending between two boundaries−the SVC posteriorly and the tricuspid annulus anteriorly in 14 of 17 cases. However, in 3 cases, activation proceeded almost simultaneously anterior and posterior to the SVC, fused around it, and then proceeded to activate the anterolateral RA wall craniocaudally. This suggests that the superior RA between the SVC and the tricuspid annulus is not necessarily an obligatory isthmus for type I atrial flutter. Activation of the medial-septal aspect of the RA (above the coronary sinus ostium) is altogether much more variable, since no consistent pattern (except for caudocranial activation) could be discerned. Although we could not corroborate previous data that the region between the coronary sinus ostium and the tricuspid valve annulus confined the circuit,17,24 our data indicate that the activation wave front widens considerably as it ascends upward in the septum. Also, the fossa ovalis is probably not associated with activation, skirting it or proceeding around what might be considered a potential endocardial obstacle. The inconsistent and somewhat irregular activation patterns in the septum could represent local “blind alleys” that may underlie previous findings of a long post pacing interval after entrainment from a fluoroscopically demarcated fossa ovalis.18 Anatomic studies showing prominent circumferential bundles encircling the anterior, lateral, and posterior aspects of the tricuspid valve with narrowing and heterogeneity of fiber arrangement in the medial and low septal regions support these activation patterns.20,21 Last, a consistently obligatory activation was observed in all cases through the IVC-TA, confirming that this structure still remains the target of choice for interrupting typical atrial flutter by transcatheter RF energy application. However this isthmus may be bypassed in some types of atypical atrial flutter circuits. Estimated conduction velocity was significantly slower (0.6 m/s) in the medial IVC-TA isthmus compared with other limbs of the circuit, where it was ≈1 m/s. There was no statistically significant difference in conduction velocity caused by amiodarone effect.
Regional Disparities of Complex Atrial Potentials During Flutter
Unlike previous studies using conventional catheter mapping4,19,22,23 the incidence and distribution of double potentials was quantitatively estimated in our study at 22% of mapped points. They were constantly present in the posterior right atrium and less frequently in the septum as well as near the coronary sinus (which includes the region corresponding to the eustachian ridge). Such potentials may be associated with the take-off site of diverging posterior right atrial wave fronts. We did sometimes find double potentials in the lateral RA wall and in the superior anterior RA but significantly less frequently and they were not associated with major changes in activation patterns probably because the changes were too circumscribed. Longer interspike double potential intervals were found in the posterior RA as compared with the double potentials in other regions in the RA. Similarly, during atrial flutter fractionated potentials were most extensive as well as invariably present in the posterior right atrium, and sometimes in a wide area. They were less frequent in the area around the tricuspid annulus probably correlating with smooth and even impulse propagation here. Although the interspike intervals would obviously vary according to the location of the recording electrode relative to the line/area of block, and the direction of propagating wave fronts, they support the role of the pericristal posterior RA as the preferential zone of conduction slowing around which the flutter circuit appears to be formed.4,13,19
The surface ECG flutter waves of patients with variant patterns on 3D EA mapping were not specifically distinctive. The lack of correlation with any distinctive surface ECG features is evidence that most of these differences in activation patterns are too subtle to be evident on the surface. Conversely the variations noted in the surface ECG flutter wave patterns may reflect differences in individual volume conductor properties and/or differences in activation in unmapped areas ie, the left atrium.
Although common atrial flutter is generally a stable arrhythmia with a stable cycle length, the software parameters used here do not allow data acquisition if consecutive local activation times vary beyond preset limits and thus provides an averaging mechanism. Although double spike electrograms are commonly considered to represent two activation fronts on either side, we selected one spike in the context of surrounding activation (to mark the local activation time). Sequential mapping resulted in a less even distribution of mapped points compared to multielectrode mapping data but when compared to the range of atrial volumes encountered, the mean density of mapping in our series was quite high. Qualitative analysis of 3D propagation was performed on software generated animation sequences with attendant limitations of interpolation, as with the analysis of traditional isochronal maps.25,26 An activation pattern indicating fusion/collision also cannot be differentiated from an underlying block between two wave fronts, which in fact may arrive locally simultaneously or nearly so, thus precluding the recording of double potentials. The assessment of regional conduction velocities was limited by the estimate of intervening curved endocardial distances. However, the estimates do provide a guide to regional differences as well as a possible index of similar drug effect. The phasic effects of respiration on the absolute intrathoracic position of the catheter tip did not contribute to the relatively wide zone of posterior double potentials in view of the corrective input from the sensor equipped and similarly positioned reference catheter. This was confirmed by the lack of artifactually wide 3D reconstructions, the precise capability of relocating known landmarks, such as the His bundle and supported by evidence from recent experimental studies.25,27
By performing panoramic mapping of the RA and using unique reconstructive software, we derived 3D activation maps during typical counterclockwise atrial flutter in humans. The circuit as revealed by high density 3D mapping indicates significant variations between individuals. The posterior limit of the circuit is marked by a ragged and wide zone of slow conduction surmounted by the orifice of the SVC in the majority. Fusion of activation wave fronts around the SVC can occur. Medial-septal RA activation though caudocranial in all does not follow a consistent pattern. However, in contrast to these variations, there is a constant obligatory activation wave front proceeding with a slower estimated conduction velocity through the narrow cavotricuspid isthmus.
We are greatly indebted to Joëlle Bassibey for helping prepare the manuscript.
- Received May 12, 1997.
- Revision received August 22, 1997.
- Accepted August 28, 1997.
- Copyright © 1997 by American Heart Association
DeBakker JMT, Hauer RNW, Simmers TA. Activation mapping: unipolar versus bipolar recording. In: Zipes D, Jalife J, ed. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. Philadelphia, Pa: WB Saunders Co; 1995:1068.
Puech P, Latour H, Grolleau R. Le flutter et ses limites. Arch Mal Coeur. 1970;63:116–144.
Chauvin M, Brechenmacher C, Voegtlin JR. Application de la cartographie endocavitaire à l’étude du flutter auriculaire. Arch Mal Coeur. 1983;76:1020–1030.
Feld GK, Fleck RP, Chen PS, Boyce K, Bahuson TD, Stein JB, Calisi CM, Ibassa M. Radiofrequency catheter ablation for the treatment of human type I atrial flutter: identification of a critical zone in the reentrant circuit by endocardial mapping techniques. Circulation. 1992;86:1233–1240.
Waldo AL. Atrial flutter: mechanisms, clinical features and management. In: Zipes, Jalife, eds. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. Philadelphia, Pa: WB Saunders Co; 1995:666.
Saoudi N, Atallah G; Kirkorian G, Touboul P. Catheter ablation of the atrial myocardium in human type I atrial flutter. Circulation. 1990;81:762–771.
Poty H, Saoudi N, Aziz AA, Nari M, Letac B. Radiofrequency catheter ablation of type I atrial flutter: prediction of late success by electrophysiological criteria. Circulation. 1995;92:1389–1392.
Cauchemez B, Haïssaguerre M, Fischer B, Thomas O, Clémenty J, Coumel P. Electrophysiological effects of catheter ablation of inferior vena cava-tricuspid annulus isthmus in common atrial flutter. Circulation. 1996;93:284–294.
Olgin J, Kalman J, Fitzpatrick A, Lesh MD. Role of right atrial structures as barriers to conduction during human type I atrial flutter: activation and entrainment mapping guided by intracardiac echocardiography. Circulation. 1995;92:1839–1848.
Shah DC, Haïssaguerre M, Jaïs P, Gencel L, Chouairi S, Clémenty J. Disparate electrical SVC activity during atrial fibrillation. Eur J Cardiac Pacing Electrophysiol. 1996;6:70. Abstract.
Cosio FG, Arribas F, Lopez-Gil M, Palacios J. Atrial flutter mapping and ablation, I: studying atrial flutter mechanisms by mapping and entrainment. PACE. 1996;19:841–853.
VanHare GF, Waldo AL. The atrial flutter reentrant circuit. Additional pieces of the puzzle. Circulation. 1996;94:244–246. Editorial.
Frame LH, Page RL, Hoffman BF. Atrial reentry around an anatomic barrier with a partially refractory excitable gap: a canine model of atrial flutter. Circ Res. 1986;58:495–511.
Kalman JM, Olgin JE, Saxon LA, Fisher WG, Lee RJ, Lesh MD. Activation and entrainment mapping defines the tricuspid annulus as the anterior barrier in typical atrial flutter. Circulation. 1996;94:398–406.
Wang K, Ho SY, Gibson DG, Anderson RH. Architecture of atrial musculature in humans. Br Heart J. 1995;73:559–565.
Nakagawa H, Lazzara R, Khastgir T, Beckman JK, McClelland JH, Imai S, Pitha JV, Becker AE, Arruda M, Gonzalez MD, Widman LE, Rome M, Neuhauser J, Wang X, Calame JD, Goudeau MD, Jackman WM. Role of the tricuspid annulus and the Eustachian valve/ridge on atrial flutter: relevance to catheter ablation of the septum isthmus and a new technique for rapid identification of ablation success. Circulation. 1996;94:407–424.
Gepstein L, Hayam G, Benttaim SA. A novel method for non fluoroscopic catheter-based electroanatomical mapping of the heart: in vitro and in vivo accuracy results. Circulation. 1997;95:1611–1622.
Kaltenbrunner W, Cardinal R, Dubuc M, Shenasa M, Nadeau R, Tremblay G, Vermeulen M, Savard P, Pagé PL. Epicardial and endocardial mapping of ventricular tachycardia in patients with myocardial infarction: is the origin of the tachycardia always subendocardially localized? Circulation. 1991;84:1058–1071.
Shpun S, Gepstein L Hayam G, Ben-Haim SA. High accuracy of radiofrequency catheter ablations guided by non-fluoroscopic mapping system in the swine atrium. J Am Coll Cardiol. 1997;29:331-A. Abstract.