Mapping and ablation of atrial tachyarrhythmias : from signal to substrate

(1)

Mapping and ablation of atrial tachyarrhythmias : from signal to

substrate

Groot, N.M.S. de

Citation

Groot, N. M. S. de. (2006, September 14). Mapping and ablation of atrial tachyarrhythmias

: from signal to substrate. Retrieved from https://hdl.handle.net/1887/4915

Version:

Corrected Publisher’s Version

License:

Licence agreement concerning inclusion of doctoral thesis in the

_{Institutional Repository of the University of Leiden}

Downloaded from:

https://hdl.handle.net/1887/4915

(2)

c

h

a

p

t

e

r

3

S - w a v e P r e d o m in a n c e o f

E p ic a r d ia l E le c t r o g r a m s d u r in g

A t r ia l F ib r illa t io n in H u m a n s :

In d ir e c t E v id e n c e fo r a R o le o f

t h e T h in S u b - E p ic a r d ia l L a y e r

Richard P.M. Houben, Natasja M.S. de Groot,

(3)

102 A b s t r a c t

Introduction: Electrograms recorded during atrial fibrillation (AF) show a high degree of spatio-temporal v ariation. C haracteriz ation of the morphology of fibrillation electrograms in patients may prov ide insight into the underly ing electro-pathological substrate of AF.

M ethods: In 2 5 patients undergoing cardiac surgery AF was induced by rapid atrial pacing. A unipolar mapping array of 2 4 4 electrodes was positioned on the free wall of the right atrium to record multiple epicardial fibrillation electrograms. L ocal anisotropy in conduction and epicardial wav efront curv ature during AF were determined by fitting the best q uadratic surface on the activ ation times of rectangular areas of 3 x 3 electrodes.

(4)

S -w a ve P re d o m in a n ce o f E p ic a rd ia l E le ct ro g ra m s d u rin g A tr ia l F ib ril la tio n in H u m a n s C h a p te r 3

103 Introduction

Despite the complex nature of atrial fibrillation (AF), the majority of unipolar electro-g rams recorded durinelectro-g acutely induced AF are of hielectro-g h amplitude w ith a sinelectro-g le neelectro-g ativ e defl ection.1_{Notches and multiple components are due to local differences in conduction}

v elocity, piv oting of w av efronts and electrical dissociation of fibrillation w av es.1_Factors

that promote fractionation include tissue anisotropy, the presence of collag eneous sep-tae, spatial dispersion in refractoriness and a hig h curv ature of w av efronts.2 -5

AF-electro-g rams thus do not only prov ide the local moments of activ ation, but they also contain important information about the underlying electro-anatomical substrate. Wells et al. used a sing le bipolar fibrillation electrog ram to disting uish different types of AF.6_{O thers}

hav e used fractionation as a marker of atrial v ulnerability.7 -9_{In the present study w e}

(5)

104 M ethods

For this study a large data set of atrial fibrillation electrograms was used as recorded from 25 patients (age 3 2 ± 11 years; 64 % male) by K onings et al.10_{All patients had normal}

non-dilated atria, no history of persistent AF and no valvular heart disease. During thoracic surgery, atrial fibrillation was induced by rapid atrial pacing. A mapping electrode con-taining 24 4 silver electrodes (diameter 3 .6 cm) was positioned on the free wall of the right atrium. The unipolar electrograms were band-pass filtered (0 .5-50 0 Hz) and sampled at 1 kHz. Before analysis the Q RS complexes were removed by template subtraction11_and

drift of the baseline was corrected by high-pass filtering (1 Hz). Electrograms containing movement artifacts were discarded. In each patient 12-20 seconds of AF were analyzed (18 .6 ± 3 .8 s) resulting in 8 0 -13 0 consecutive fibrillation maps. The RS-difference of the fibrillation electrograms was calculated as the difference in R and S amplitude divided by the total amplitude of the fibrillation electrogram:

(1)

Quantification of Anisotropy during Atrial Fibrillation

Based on the 24 4 simultaneously recorded electrograms, isochrone maps were construct-ed for each episode of AF (8 7 ± 3 5 maps per patient). The fibrillation maps were usconstruct-ed to measure the local conduction vectors in areas of 4 .5 x 4 .5 mm (3 x 3 electrodes). A q ua-dratic surface was fitted to the set of 9 activation times by the following formula:

(2) where t_{ac t} is the activation time at each of the 9 recording sites at the coordinates x and y. The coefficients a₁ to a₆, describing the fitted q uadratic surface, can be calculated if at least 6 of the 9 activation times are available. The curvature of the wavefront at the center of the area of 4 .5 x 4 .5 mm is defined by the inverse radius of the circle fitted to the wavefront:

(3)

where K is the curvature (cm– 1_{) and x and y all coordinates within the fitted q uadratic}

surface. The conduction velocity and direction are given by the maximum gradient of the q uadratic surface at the center of the 4 .5 x 4 .5 mm area. The directional differences in

(6)

105

conduction velocity during AF were quantified as illustrated in Figure 1. For each fibril-lation wave the local conduction vectors were calculated in areas of 4.5 x 4.5 mm (left panel). V ectors with a velocity of less than 7.5 cm/ s (conduction block) were excluded. Local anisotropy in conduction was measured by fitting the best ellipse through a large set of conduction vectors of fibrillation waves passing under the electrode during a period of 20 seconds of AF. Within this time frame the fibrillation waves traveled in all differ-ent directions and with differdiffer-ent curvatures. The ellipse was described by the parametric equation:12

(4)

where V_x and V_y is a projection of the set of velocity vectors on two perpendicular axes. The ellipse is completely determined by a and b related to the length of the long and short axes and the rotation angle θ. The polar angles are given by φ ; 0… 2π.. Since the ellipse was fitted in a least squares sense, given a large enough dataset, the influence of other factors than anisotropy averaged out.13_{The rotation angle was obtained by an iterative search of}

V_x V_y

=

a co s(θ) – a sin (θ)

b sin (θ) b co s(θ)

co s(φ) sin (φ)

, Figure 1.

(7)

106

the most distal point on the ellipse while varying the polar angle. When the rotation angle is given, the long and short axes of the ellipse can be solved from the function prescrip-tion. The ratio between the long and short axis of the best fitted ellipse was taken as a measure of the local anisotropy in conduction (Figure 1). This method has been validated by pacing from four different directions and comparing the rotation angle of the fitted el-lipse with the direction of the major pectinate muscles obtained by macro-photography.14

A correlation coefficient of r = 0.77 (p < 0.001) was found with a mean error of less than 2 degrees.

Computer Simulations

To investigate the relationship between wavefront curvature, transmural dispersion and RS-difference, computer simulations were performed using the boundary element method.15

Infinite medium potentials were calculated from a uniform double layer at each position of the activation wave in a homogeneous isotropic medium of 50x50x2 mm. A set of concentric activation waves was placed in this medium starting from one of the edges. By definition, the inverse radius of each circular wavefront is defined by its curvature. To investigate the influence of different degrees of epi- to endocardial activation, the doubles layers were placed at different angles. The transmural angle was kept constant through-out the whole medium. The potential at each site in the medium is proportional to the solid angle subtended by the double layer. The numerical solution of the solid angle is found by triangulation of the double layer. In this way the electrograms of the epicardial surface were constructed.

Electrograms close to the origin of the wave are associated with a high wavefront curva-ture, whereas at more remote sites the electrograms are generated by planar wavefronts. The effects of differences in epi- and endocardial activation on R and S wave amplitude

were measured by reconstructing epicardial electrograms during simulation of wavefronts at different transmural angles.

Statistical Analysis

(8)

107 R esults

Anisotropy in Conduction during AF

In 25 patients the lateral mid-wall of the right atrium was mapped (244 electrodes) dur-ing acutely induced atrial fibrillation. C ontinuous segments of 18.6 ± 3.8 seconds of AF were analyzed, resulting in a total database of 2,226 maps and 413,031 unipolar fibril-lation potentials. In Figures 2 and 3, fifteen consecutive fibrilfibril-lation maps are shown of two different patients. In one patient, activation of the right atrium was more organized than in the other. In Figure 2 (patient # 10) the mean fibrillation interval was 134 ± 17 ms. Most of the maps show two fusing or colliding wavefronts entering the mapping area from different directions (arrows). Frequently, the right atrium was activated by only one fibrillation wave (maps 6, 7, 9, 11, 14 and 15). Due to this uniform activation the effective conduction velocity of the fibrillation waves could be easily determined. Only minor directional differences in conduction velocity of the fibrillation waves occurred. For instance, two fibrillation waves propagating in perpendicular directions (beats 6 and 11) showed an effective conduction velocity of respectively 54 and 52 cm/s (anisotropy ratio 1.04). Selecting the two perpendicular wavefronts with the largest difference in conduc-tion velocity (beats 7 and 15) still yielded an anisotropy ratio of only 1.15. In this patient anisotropy mapping in areas of 4.5x4.5 mm resulted in a median anisotropy ratio of 1.19. In the other patient shown in Figure 3 the right atrium was activated in a more complex way. Although the mean fibrillation interval (159 ± 31 ms) was actually longer than in the patient of Figure 2, activation was less organized. Almost all maps showed multiple wavefronts separated by lines of functional conduction block. In addition the fibrillation waves not only entered the mapping area from outside, but also appeared as epicardial breakthrough under the mapping electrode (asterisks). Frequently wavefronts pivoted around the end of a line of block and made a full 180° turn. C losed loop reentry was sometimes observed but did not occur during the 15 maps shown in Figure 3. A stable reentrant circuit was not seen. Due to this more disorganized type of fibrillation it was more difficult to measure conduction velocity. Y et, some of the fibrillation waves were broad enough and traveled over a sufficient distance to allow reliable measurement of their effective conduction velocity. The vectors of these large fibrillation waves yielded an anisotropy ratio of 1.21. At a sub-macroscopic scale (areas of 4.5 x 4.5 mm) the median anisotropy ratio was 1.18.

RS-difference of U nipolar Fibrillation Electrograms

(9)

108

profound (mean RS-difference -0.15 and -0.22 respectively). In type III AF the prevalence of S-waves was less prominent (-0.07; p<0.005). The incidence of fibrillation potentials with an almost complete S-wave morphology (RS-difference < -0.5) was 5.3% ± 3.7 and 11.8% ± 7.2 during type I and II AF, compared to 3.4% ± 1.6 during type III AF (p<0.05). Opposite, R-waves (RS-difference > 0.5) were more rare during type I and II AF (0.4% ± 0.3 and 0.5% ± 0.1) than in type III AF (1.1% ± 0.4; p<0.005).

In Table I the 25 patients are ranked according to the complexity of AF, using the percent-age of local conduction block (velocity<7.5 cm/s) as an index for the degree of electrical dissociation. Less then 5% of conduction block was assigned as type I, between 5 and 10% as type II, and > 10% as type III AF. On average, the local conduction velocity of fibrillation waves was lower (p<0.05) during type III AF (47.8 ± 7.2 cm/s) compared to type II (58.3 ± 3.9 cm/s) and type I AF (67.4 ± 8.9 cm/s). Comparing conduction in perpendicular

Figure 2.

(10)

109

directions with the highest and lowest velocity yielded a mean ‘longitudinal’ conduction velocity of 62.6 ± 9.1 and a ‘transverse’ velocity of 50.5 ± 7.9 cm/s (median anisotropy ratio 1.24 ± 0.09). No difference in anisotropy ratio between the three types of AF was found (p>0.65). Even taking the highest degree of directional differences in conduction (p₉₅) the anisotropy ratio during AF was still not more than 1.80 ± 0.39.

RS-difference and Anisotropy

Slow conduction transverse to the fiber axis is associated with electrograms of low am-plitude and a predominance of S-waves.2_{To test whether the observed negative}

RS-dif-ference was due to transverse conduction, we calculated the RS-difRS-dif-ference during con-duction perpendicular to the long anisotropy axis (‘transverse’ concon-duction). The mean RS-difference in this sub-group was -0.15 ± 0.08 during type I, -0.22 ± 0.07 during type

Figure 3.

(11)

(12)

111

II, and -0.07 ± 0.05 during type III AF (Table I). The mean RS-difference recorded during conduction parallel to the axis of anisotropy (‘longitudinal’ conduction) was -0.14 ± 0.08 during type I, -0.23 ± 0.08 during type II and -0.07 ± 0.05 during type III. No statistical difference in RS-difference was found between electrograms recorded during ‘transverse’ and ‘longitudinal’ conduction (p>0.88). The predominance of S-waves in the epicardial fibrillation electrograms thus could not be explained by anisotropy in conduction. RS-difference and Wavefront Curvature

Wavefront curvature simulation

The effect of wavefront curvature on RS-difference was investigated by computer simu-lations in a sheet of isotropic tissue of 5x5 cm (upper panels Figure 5). Electrograms calculated for convex and concave wavefronts (radius between 0.1 and 5 cm) revealed an S-shaped relationship between curvature and RS-difference; convex wavefronts produced an rS and concave wavefronts an Rs morphology. When a convex wavefront approaches a recording electrode the solid angle ‘seen’ by the electrode is smaller than in case of a planar wavefront. In contrast, after a curved wavefront has passed the electrode the solid

Figure 4.

(13)

112

angle becomes wider, resulting in a small and narrow r-wave followed by a large and wide S-wave. Activation waves with a radius of <0.1 cm resulted in almost pure S-waves (RS-difference<-0.89).

Simulation of epi to endocardial propagation

The effects of differences in endo-epicardial activation were determined by simulating wavefronts propagating with a different transmural angle (lower panels Figure 5). Early activation of the endocardium produced epicardial potentials with pronounced R-waves. Opposite, when the epicardium was activated first, electrograms with an rS morphology were generated. The lower right panel of Figure 5 shows the quantitative relationship be-tween the transmural angle of activation and the RS-difference of epicardial electrograms.

Figure 5.

To determine the relationship between wavefront curvature and epi- to endocardial activation on the difference in relative RS-amplitude, computer simulations were performed using a model of a 50 x 50 x 2 mm sheet of iso-tropic tissue. Upper panels: Five concentric activation waves are shown originating from the upper-mid location of the sheet. Simulated electrograms are shown with each activation wave. An S-shaped relationship was found between wavefront curvature and RS-difference. Convex wavefronts (curvature < 0) revealed more prominent S-waves whereas concave wavefronts produced an Rs morphology.

(14)

113

Oblique wavefronts of 45° resulted in RS-differences of + or -0.51. R or S waves were produced by transmural wavefronts with an angle of >75o_.

Measurement of Wavefront Curvature during AF

To determine whether the predominance of epicardial S waves was due to a high curvature of the fibrillation waves, in all 2,226 fibrillation maps the local curvatures were correlated with the RS-differences. High curvatures existed around sites of epicardial breakthrough and at pivot points. Concave wavefronts resulted from fusion of fibrillation waves. In Figure 6 the relationship between curvature and RS-difference during AF is given for all 25 patients. In the lower left panel the distribution of wavefront curvature is shown. Only a slight prevalence of convexity was found (median curvature -0.2 ± 0.1 cm-1_{). The}

in-Figure 6.

Measurement of the curvature of fibrillation waves in the free wall of the human right atrium. In the three maps examples of convex, planar and concave fibrillation waves are shown. In the lower left panel the histogram of all

curvatures is plotted (n = 413,031; 25 patients). Curvatures with a K value < -2.5 cm-1_{were assigned as convex}

and curvatures >2.5 cm-1 _{as concave. In the lower right panel the curvature is correlated with the RS-difference}

(15)

114

cidence of either highly convex (K <-2.5 cm-1_{) or concave wavefronts (K >2.5 cm}-1_{) was}

similar (11.8% ± 3.9 and 11.1% ± 3.7). Regression analysis of wavefront curvature and RS-difference showed a positive trend (r = 0.23; p <0.01). The correlation coefficients between curvature and RS-difference for “ longitudinal” and “ transverse“ propagation were 0.24 and 0.12 (p <0.01). Since the average RS-difference was negative at all cur-vatures, another factor must be involved in causing epicardial potentials with an S-wave morphology.

Right Atrial Anatomy

The right atrium consists of a smooth walled compartment and a highly trabeculated part. The former is known as the intercaval area, which also contains the interatrial septal surface and is delineated by the oval fossa. The intercaval compartment is separated from the trabeculated part, also known as the right atrial appendage or auricle, by the terminal crest. From the terminal crest the pectinate muscles originate, at almost perpen-dicular angles, thus accounting for the trabeculated nature of the right atrial appendage. The architecture of the right atrial appendage, therefore, is characterized by myocardial

trabeculae running in almost parallel fashion (albeit with multiple crossovers) connected by a thin, almost translucent layer of atrial myocardium. As shown in Figure 7 this thin layer is composed of myocardial cells which are in direct contiguity with the underlying trabeculae.

Figure 7.

(16)

115 Discussion

Measurement of Anisotropy during AF

The architecture of the atria is highly anisotropic, comprising trabeculated areas and muscle bundles with fibers running in parallel. Indeed, in isolated atrial preparations a high anisotropy ratio in conduction (between 3 and 5) has been measured.2,16_{In contrast}

however, on a macroscopic scale the atria revealed only minor directional differences in conduction velocity.17,18_{To evaluate the atrial anisotropy during fibrillation, we measured}

the directional differences in conduction velocity in areas of 4.5x4.5 mm. Since during atrial fibrillation the activation waves change constantly in direction, mapping of 12-20 seconds of AF provides a map of omni-directional conduction vectors. This allowed accu-rate measurement of the degree of anisotropy in conduction during AF. In the total group of 25 patients the median anisotropy ratio in the right atrium ranged between 1.13 and 1.52 (1.24 ± 0.09). This remarkably low value is in agreement with the study of Hansson

et al.17_{but is far less than the anisotropy ratio measured in small preparations.}2,16

Predominance of S-waves in Epicardial Fibrillation Electrograms

The unipolar fibrillation electrograms recorded from the epicardium of the right atrium showed a clear predominance of S waves. Especially in patients in whom right atrial ac-tivation during AF was relatively uniform, a high predominance of S-waves was found. In type I and II AF the mean RS-differences were respectively -0.15 ± 0.08 and -0.22 ± 0.08 with an incidence of S waves (RS-difference < -0.5) of more than 7.5%. In contrast, pa-tients with more ‘disorganized’ AF (type III) showed less S-wave prevalence with a mean RS-difference close to zero (-0.07 ± 0.05; p<0.005) and an incidence of S waves less than 3.5% (p<0.05).

There are several possible explanations for the predominance of S-waves in epicardial elec-trograms: 1) Tissue anisotropy and conduction transverse to the fiber orientation2,19_{2) A}

high curvature of fibrillation waves4,20_{3) Preferential epi- to endocardial conduction.}17,21

Tissue anisotropy creates spatial differences in electrogram amplitude and morphology.22

In uniform anisotropic tissue, low amplitude S-waves are recorded from locations where the impulse propagates transverse to the fiber axis.2_{We correlated the RS-difference with}

the direction of propagation of fibrillation waves. The RS-difference recorded during con-duction along the long axis of anisotropy (‘longitudinal’ concon-duction) was compared with electrograms during ‘transverse’ conduction. No significant difference in RS-difference was found during ‘transverse’ or ‘longitudinal’ conduction. The low degree of anisotropy itself, together with the absence of a relationship between S-waves and ‘transverse’ con-duction, thus makes it unlikely that tissue anisotropy plays a major role in the origin of the RS-differences.

(17)

116

is smaller when it approaches the recording electrode than when it is propagating away from it. In our computer simulations, variation of the wavefront curvature between –10 cm-1_{(convex) and +10 cm}-1 _{(concave) resulted in RS-differences between -0.89 and +0.89.}

In reality the maximal convexity of a wavefront is limited, since above a critical value of 50 cm-1_{conduction will fail.}4,24_{In our database of 413,031 fibrillation potentials no}

statistically significant correlation was found between the relative amplitude of the R and S-waves and the epicardial wavefront curvature (r = 0.23; p<0.01). In addition, potentials with an rS morphology were frequently recorded during propagation of planar wavefronts. Therefore, the convex curvature of fibrillation waves can not explain the predominance of

S-waves in epicardial electrograms.

Epi- to endocardial activation of the atrial wall generates epicardial electrograms with an rS morphology.21_{In the present study we did not record electrograms from the}

endocardi-um and therefore a direct evaluation of the role of transmural activation of the atrial wall during AF was not possible. Computer simulations revealed a prominent role of a leading epicardial wavefront in generating epicardial electrograms with a negative RS-difference. When the transmural angle was >75o_{the electrograms recorded from the epicardium}

showed an almost pure S-wave morphology. Epi-endocardial activation mapping of iso-lated canine atria during sinus rhythm revealed local transmural differences in activation up to 13 ms.21_{In the thicker human atria differences in epi-endocardial activation time}

can even be higher. Since anisotropy in conduction and wavefront curvature could not explain the observed negative RS-difference in epicardial fibrillation electrograms, we sug-gest that the S-wave predominance is a sign of preferential epi- to endocardial activation during atrial fibrillation.

A Proposed Role of the Thin Epicardial Layer

A leading role of the sub-epicardial layer for activation of the atrial wall during sinus rhythm and atrial pacing has been demonstrated by Schuessler et al.21_{We propose that}

the thin epicardial layer of atrial myocardium plays an important role in propagation of fibrillation waves. In this respect it is of interest that in patients with a more complex type of AF (type III), the predominance of S-waves was significantly less then in patients with type I and II AF. This may not only be due to a higher curvature of the fibrillation waves during type III AF, but also to a loss of the ‘leading role’ of the epicardial layer. In case of electrical discontinuities in the thin sub-epicardial sheet of atrial myocytes, the trabeculated architecture of the atria will determine the pathways taken by the fibrillation waves. This will lead to a more disorganized type of AF. For a more detailed understand-ing of the role of the epicardial layer simultaneous epi- and endocardial recordunderstand-ings dur-ing atrial fibrillation are necessary. Comparison of endo- and epicardial RS-morphology will give more insight in the 3-D activation of the atrial wall. So far, most studies have been done either with endocardial or epicardial electrodes.10,25-29_{Only a few studies allow}

(18)

117

et al. showed that the epi- and endocardium of the canine right atrium were activated asynchronously during atrial arrhythmias.21_{Particularly in thicker parts of the atrial wall}

the epicardium was activated before the endocardium, and endocardial electrograms recorded from pectinate muscles showed more prominent R-waves compared to the epicardium. In contrast, endocardial electrograms recorded between pectinate muscles were activated at about the same time as the epicardium and showed a similar morphol-ogy. In Figure 8 the theoretical leading role of the thin epicardial layer is illustrated. An epicardial wavefront is propagated uniformly transverse to the orientation of a number of endocardial trabeculae. The crest of the epicardial wavefront activates the trabeculae in epi- to endocardial direction. From these earliest points of activation the impulse will propagate longitudinally along the long axis of the pectinate muscles. At the epicardium, due to the epi- to endocardial direction of activation, the unipolar electrograms will show a predominant wave morphology. At the endocardium the amplitude of the R- and S-waves will be variable depending on the distance from the earliest point of activation of the trabeculae. Due to the complex structure of the atrial wall, the morphology of an en-docardial electrogram thus is not necessarily a mirror image of the electrogram recorded

Figure 8.

(19)

118

at the epicardium. Quantitative evaluation of endocardial electrograms during atrial fi-brillation requires a large number of electrograms. Non-contact mapping techniques can generate such large data sets of endocardial electrograms during human AF. Comparison between contact and non-contact mapping of the right atrium demonstrated that the computed endocardial signals showed a great similarity in morphology with directly re-corded electrograms.30

Study Limitations

This study is based on the analysis of non-fractionated potentials during acutely induced AF in a group of young patients without a history of AF. Only the right atrium was mapped, whereas clinically the pathophysiological substrate of AF is often located in the left atrium. The present data thus may not be representative for patients with persistent AF. Another

(20)

119 References

1. Konings, K. T., Smeets, J. L., Penn, O. C., Wellens, H. J., and Allessie, M. A. Configuration of polar Atrial Electrograms During Electrically Induced Atrial Fibrillation in Humans. Circulation

3-4-1997;95(5):1231-41.

2. Spach, M. S. and Dolber, P. C. Relating Extracellular Potentials and Their Derivatives to Anisotropic Propagation at a Microscopic Level in Human Cardiac Muscle. Evidence for Electrical Uncoupling of Side-to-Side Fiber Connections With Increasing Age. Circ.Res. 1986;58(3):356-71.

3. Huang, J. L., Tai, C. T., Chen, J. T., Ting, C. T., Chen, Y. T., Chang, M. S., and Chen, S. A. Effect of Atrial Dilatation on Electrophysiologic Properties and Inducibility of Atrial Fibrillation. Basic Res. Cardiol. 2003;98(1):16-24.

4. Fast, V. G. and Kleber, A. G. Role of Wavefront Curvature in Propagation of Cardiac Impulse. Cardio-vasc.Res. 1997;33(2):258-71.

5. Ndrepepa, G., Caref, E. B., Yin, H., el Sherif, N., and Restivo, M. Activation Time Determination by High-Resolution Unipolar and Bipolar Extracellular Electrograms in the Canine Heart. J.Cardiovasc. Electrophysiol. 1995;6(3):174-88.

6. Wells, J. L., Jr., Karp, R. B., Kouchoukos, N. T., MacLean, W. A., James, T. N., and Waldo, A. L. Characterization of Atrial Fibrillation in Man: Studies Following Open Heart Surgery. Pacing Clin. Electrophysiol. 1978;1(4):426-38.

7. Ohe, T., Matsuhisa, M., Kamakura, S., Yamada, J., Sato, I., Nakajima, K., and Shimomura, K. Relation Between the Widening of the Fragmented Atrial Activity Z one and Atrial Fibrillation. Am.J.Cardiol. 12-1-1983;52(10):1219-22.

8. Tanigawa, M., Fukatani, M., Konoe, A., Isomoto, S., Kadena, M., and Hashiba, K. Prolonged and Fractionated Right Atrial Electrograms During Sinus Rhythm in Patients With Paroxysmal Atrial Fi-brillation and Sick Sinus Node Syndrome. J.Am.Coll.Cardiol. 1991;17(2):403-8.

9. Tai, C. T., Chen, S. A., Tzeng, J. W., Kuo, B. I., Ding, Y. A., Chang, M. S., and Shyu, L. Y. Prolonged Fractionation of Paced Right Atrial Electrograms in Patients With Atrial Flutter and Fibrillation. J.Am. Coll.Cardiol. 2001;37(6):1651-7.

10. Konings, K. T., Kirchhof, C. J., Smeets, J. R., Wellens, H. J., Penn, O. C., and Allessie, M. A. High-Den-sity Mapping of Electrically Induced Atrial Fibrillation in Humans. Circulation 1994;89(4):1665-80. 11. Hoekstra, B. P., Diks, C. G., Allessie, M. A., and Degoede, J. Non-Linear Time Series Analysis:

Meth-ods and Applications to Atrial Fibrillation. Ann.Ist.Super.Sanita 2001;37(3):325-33. 12. Weisstein, E. W., The CRC Concise encyclopedia of Mathematics. 1999.

13. Gander W, Golub G.H., and Strebel R. Fitting of Circles and and Ellipses: Least Square Solution. BIT 1994;34:556-77.

14. Eijsbouts, S. C. M., Houben, R. P. M., Blaauw, Y., Schotten, U., and Allessie, M. A. Synergistic Ac-tion of Atrial DilataAc-tion and Sodium Channel Blockade on ConducAc-tion in Rabbit Atria. (Manuscript under Review) 2004.

15. Van Oosterom, A. Solidifying the Solid Angle. J.Electrocardiol. 2002;35 Suppl:181-92.

16. Saffitz, J. E., Kanter, H. L., Green, K. G., Tolley, T. K., and Beyer, E. C. Tissue-Specific Determi-nants of Anisotropic Conduction Velocity in Canine Atrial and Ventricular Myocardium. Circ.Res. 1994;74(6):1065-70.

17. Hansson, A., Holm, M., Blomstrom, P., Johansson, R., Luhrs, C., Brandt, J., and Olsson, S. B. Right Atrial Free Wall Conduction Velocity and Degree of Anisotropy in Patients With Stable Sinus Rhythm Studied During Open Heart Surgery. Eur.Heart J. 1998;19(2):293-300.

18. Koura, T., Hara, M., Takeuchi, S., Ota, K., Okada, Y., Miyoshi, S., Watanabe, A., Shiraiwa, K., Mita-mura, H., Kodama, I., and Ogawa, S. Anisotropic Conduction Properties in Canine Atria Analyzed by High-Resolution Optical Mapping: Preferential Direction of Conduction Block Changes From Longitudinal to Transverse With Increasing Age. Circulation 4-30-2002;105(17):2092-8.

19. Corbin, L. V. and Scher, A. M. The Canine Heart As an Electrocardiographic Generator. Dependence on Cardiac Cell Orientation. Circ.Res. 1977;41(1):58-67.

(21)

120

21. Schuessler, R. B., Kawamoto, T., Hand, D. E., Mitsuno, M., Bromberg, B. I., Cox, J. L., and Boineau, J. P. Simultaneous Epicardial and Endocardial Activation Sequence Mapping in the Isolated Canine Right Atrium. Circulation 1993;88(1):250-63.

22. Spach, M. S., Miller, W. T., III, Miller-Jones, E., Warren, R. B., and Barr, R. C. Extracellular Potentials Related to Intracellular Action Potentials During Impulse Conduction in Anisotropic Canine Cardiac Muscle. Circ.Res. 1979;45(2):188-204.

23. Simms, H. D., Jr. and Geselowitz, D. B. Computation of Heart Surface Potentials Using the Surface Source Model. J.Cardiovasc.Electrophysiol. 1995;6(7):522-31.

24. Lindemans, F. W. and Denier Van der Gon JJ. Current Thresholds and Liminal Size in Excitation of Heart Muscle. Cardiovasc.Res. 1978;12(8):477-85.

25. Sun, H., Velipasaoglu, E. O., Wu, D. E., Kopelen, H. A., Zoghbi, W. A., Spencer, W. H., III, and Khoury, D. S. Simultaneous Multisite Mapping of the Right and the Left Atrial Septum in the Canine Intact Beating Heart. Circulation 7-20-1999;100(3):312-9.

26. Lin, J. L., Lai, L. P., Tseng, Y. Z., Lien, W. P., and Huang, S. K. Global Distribution of Atrial Ectopic Foci Triggering Recurrence of Atrial Tachyarrhythmia After Electrical Cardioversion of Long-Standing Atrial Fibrillation: a Bi-Atrial Basket Mapping Study. J.Am.Coll.Cardiol. 3-1-2001;37(3):904-10. 27. Allessie MA; Lammers WJ; Smeets JR; Bonke FI. Total mapping of atrial excitation during

acetylcho-line-induced atrial flutter and fibrillation in the isolated canine heart. Kulbertus HE, Olsson SB, and SchlepprerM. Atrial Fibrillation. Molndal Sweden: Lindgren and Soner; 1982. pp.44-62.

28. Cox, J. L., Canavan, T. E., Schuessler, R. B., Cain, M. E., Lindsay, B. D., Stone, C., Smith, P. K., Corr, P. B., and Boineau, J. P. The Surgical Treatment of Atrial Fibrillation. II. Intraoperative Electrophysi-ologic Mapping and Description of the ElectrophysiElectrophysi-ologic Basis of Atrial Flutter and Atrial Fibrilla-tion. J.Thorac.Cardiovasc.Surg. 1991;101(3):406-26.

29. Kirchhof, C., Chorro, F., Scheffer, G. J., Brugada, J., Konings, K., Zetelaki, Z., and Allessie, M. Re-gional Entrainment of Atrial Fibrillation Studied by High-Resolution Mapping in Open-Chest Dogs. Circulation 1993;88(2):736-49.

(22)

S-wave Predominance of Epicardial Electrograms during Atrial Fibrillation in Humans

12

1

(23)

Mapping and ablation of atrial tachyarrhythmias : from signal to substrate