The sawtooth EKG pattern of typical atrial flutter is not related to slow conduction velocity at the cavotricuspid isthmus

Arunashis Sau, MBBS 1,2 *Markus B Sikkel, PhD 1,2 *Vishal Luther, MBBS 1,2

Ian Wright, BSc2

Fernando Guerrero, BSc 3

Michael Koa-Wing, PhD 2

David Lefroy, MBBS 2

Nicholas Linton, PhD 1,2

Norman Qureshi, PhD 2

Zachary Whinnett, PhD 1,2

Phang Boon Lim, PhD 1,2

Prapa Kanagaratnam, PhD 1,2

Nicholas S Peters, MD 1,2

D Wyn Davies, MD 1,2

 * These authors contributed equally to this work

1Imperial College London, London, United Kingdom2Imperial College Healthcare NHS Trust, Department of Cardiology, London, United Kingdom3Boston Scientific, Breakspear Park, Hemel Hempstead, Herts, United Kingdom

CORRESPONDING AUTHOR:Dr Markus B Sikkel, Myocardial Function Section, Fourth Floor, Imperial Centre for Translational and Experimental Medicine, Imperial College London, Hammersmith Campus, Du Cane Road, London W12 0NN, UK Tel: +44 207 594 2734; fax: +44 207 351 8145, Email: m.sikkel@imperial.ac.uk

Word Count: 4452

Key words: Atrial flutter, rhythmia, cavotricuspid isthmus, conduction velocity,

sawtooth

Disclosures

PBL declares receipt of a research grant and speaker’s fees from Boston Scientific. FG is an employee of Boston Scientific. The other authors have no conflicts to report.

1

Abstract

Introduction

We hypothesised that very high density mapping of typical atrial flutter (AFL) would

facilitate a more complete understanding of its circuit. Such very high density

mapping was performed with the Rhythmia mapping system using its 64 electrode

basket catheter.

Methods and Results

Data were acquired from 13 patients in AFL. Functional anatomy of the right atrium

(RA) was readily identified during mapping including the Crista Terminalis and

Eustachian ridge. The leading edge of the activation wavefront was identified without

interruption and its conduction velocity (CV) calculated. CV was not different at the

cavotricuspid isthmus (CTI) compared to the remainder of the RA (1.02 vs. 1.03 m/s,

p = 0.93). The sawtooth pattern of the surface EKG flutter waves were compared to

the position of the dominant wavefront. The downslope of the surface EKG flutter

waves represented on average, 73% ± 9% of the total flutter cycle length. During the

downslope the activation wavefront travelled significantly further than during the

upslope (182 ± 21ms vs. 68 ± 29ms, p<0.0001) with no change in conduction

velocity between the two phases (0.88 vs. 0.91 m/s, p=0.79).

Conclusion

CV at the CTI is not slower than other RA regions during typical AFL. The gradual

downslope of the sawtooth EKG is not due to slow conduction at the CTI suggesting

that success of ablation at this site relates to anatomical properties rather than

presence of a “slow isthmus”.

2

Introduction

Typical, counter clockwise (CCW) atrial flutter (AFL) has been classically described

as cavotricuspid isthmus (CTI) dependent, with the CTI region described as the floor

of the right atrium (RA) between the inferior tricuspid annulus and the inferior vena

cava1. An area of slow conduction in the low right atrium has previously been

described, later determined to be the CTI region 2-6 . The P waves during typical AFL

have a classic saw tooth pattern on the EKG, with predominantly negative

deflections in the inferior leads. One of several possible explanations for this pattern

is slow conduction through the CTI region giving a gradual downslope, followed by

fast conduction in a caudo-cranial direction along the septal wall, seen as the sharp

upslope component of the sawtooth.1 Other possible mechanisms include left atrial

and inter-atrial septum activation7.

The RhythmiaTM mapping system (Boston Scientific) offers very high mapping

density8,9. Current 3D electroanatomical mapping systems can achieve high density

mapping, however this is often more time consuming and of a lower point density.

Rhythmia circumvents these issues using a small 64-electrode basket array catheter

(“OrionTM”). The low-impedance electrodes have a small surface area (0.4mm2)

making them highly sensitive in the detection of even the smallest bipolar potentials

(e.g. in the pulmonary veins)10,11. Mapping allows localization of anatomy and

electrograms to 1mm resolution.

3

We hypothesised that high-density mapping of typical AFL would allow us to study in

more detail the relationship between the flutter circuit and conduction in the CTI

region.

4

Methods

This single centre study recruited patients undergoing AFL ablation who were

mapped with the RhythmiaTM mapping system. The Institutional Committee on

Human Research at our centre approved the study.

Procedure

Femoral venous access was attained using two 7F venous sheaths and one 9F

sheath. The patients were then fully heparinized to maintain an ACT of >300 secs

throughout the procedures. A decapolar catheter was positioned in the coronary

sinus (CS) and the Orion catheter then inserted through the 9F sheath into the right

atrium.

Full details of the RhythmiaTM system, Orion catheter and the system’s electrogram

annotation have been previously described10,12. Briefly, the Orion catheter is a

minibasket catheter with 64 small (0.4mm2) low impedance electrodes arranged in 8

splines. Mapping with the Orion catheter is performed sequentially relative to timing

of a reference catheter positioned in CS. Roving the catheter allows progressive

construction of the anatomic and activation maps. While the Orion catheter is not

specifically designed for right atrial mapping, the smooth contours of the right atrium,

in our experienced, allowed excellent geometry and activation maps to be created.

The original study validating the high density mapping of the Rhythmia system in

vivo was carried out in canine right atria. 10

5

A right atrial map was created prior to ablation. The geometry of the cardiac

chambers was acquired based on the location of the outer-most electrodes of the

basket using in-built magnetic and impedance sensing. Following mapping, CTI

ablation was performed with the operator’s preferred ablation catheter with the

assistance of impedance based catheter localization using RhythmiaTM. Points were

automatically and continuously annotated based on pre-assigned beat acceptance

criteria (cycle length within 10 ms, propagation reference within 5 ms, motion within

1mm, EGM stability of 75% and respiration within 5µV) with no further adjustment

made while mapping.

Mapping

The system calculated the median tachycardia cycle length over 10sec and set the

duration of the window of interest (WOI) to 100% of the observed cycle length,

centered on the timing reference electrode. The WOI was represented as a color

wheel running from red to purple, and spanning the colors of the rainbow. These

limits defaulted to red being the earliest in the window and purple being the latest.

These limits could be rotated around the color wheel, and this had the equivalent

effect of sliding the WOI without causing a full map re-compute. The annotation of

local activation time was assigned automatically by the system, based on the

maximum absolute peak of the bipolar electrogram. For electrograms with more than

1 potential, annotation was guided by the relative timings of surrounding

electrograms. Individual electrograms and their annotation of activation timing could

be studied by roving a virtual probe incorporated within system at the desired

location.

6

Analysis

Analysis of the maps was performed “offline”. Each map was examined with

particular attention paid to conduction velocities and areas of apparent conduction

delay and/or block. These areas underwent detailed manual inspection of

electrograms. Procedural data was collected from RhythmiaTM and LabSystem

PROTM (Boston Scientific). Analysis of conduction velocity was performed using 2

methods. For both methods, the activation time of the wavefront was recorded using

the RhythmiaTM “wheel” (a circular interface within the software allowing the user to

move the wavefront to a specific point in the tachycardia cycle). Firstly, gross

conduction velocity was calculated. To do this the tricuspid annulus was split into

hours of the clock face, which were then combined into 4 groups of 3, defining the 4

to 7 o’clock region as the CTI. The other 3 regions were the superior RA, inter-atrial

septum and lateral wall. Wavefront activation was followed by rotating the color

wheel and tracking the leading edge of the wavefront around the cardiac chamber.

The distances travelled by the dominant wavefront path were measured. The path of

the wavefront was determined by advancing the RhythmiaTM wheel in steps. The

path was then measured using the software ruler tool (a tool allowing measurement

of length in mm along the contour of the RA), and the CV calculated. We then

obtained higher resolution measures of the CV at 8 different points around the

tricuspid valve annulus (TVA) corresponding to compass points (or 12:00, 1:30, 3:00,

4:30, 6:00, 7:30, 9:00, 10:30 on the clock-face). At the centre of each compass point,

when looking at the TVA in a left anterior oblique (LAO) view, the distance travelled

by the dominant wavefront over 5ms was measured, and subsequently the CV

determined (Figure 1 in the Data Supplement).

7

The location of the wavefront was compared to the start and end of both the

downslope and upslope of the sawtooth flutter wave in EKG lead III. The time spent

in the downslope vs. upslope of the sawtooth wave was measured and the

percentage time spent in each phase calculated. The start of the downslope was

taken as the point where the upslope ended. The downslope is depicted as the red

and orange lines in Figure 1, with the green line indicating upstroke. Distance

travelled by the wavefront in each downslope and upslope was also measured using

the software ruler tool, in order to calculate CV in each phase. This was achieved by

measuring the path of the leading edge of the dominant wavefront, regardless of

where this was in relation to the TVA.

Statistical analysis was performed using paired two-tailed Student’s t-tests for

normally distributed data. One-way repeated measures ANOVA was used for

comparison of multiple parameters.

8

Results

19 patients underwent mapping of the RA and ablation of AFL (89% male, mean age

68 ± 7).

13 patients were in AFL at the start of the procedure. 10 patients were in

counterclockwise (CCW) flutter, with the remainder in clockwise (CW) flutter. Patient

demographics and basic procedural data are shown in Table. 1.

An area of conduction block at the posterior aspect of the CTI between the IVC

running towards the CS os was identified by mapping wavefronts and confirmed by

the presence of electrograms with double potentials in this region in 6 (5 with CCW,

1 with CW flutter) patients. Anatomically this area was consistent with the Eustachian

ridge (ER), as shown in Figure 2A. The average distance between TVA and IVC was

37.3 ± 7.7mm, although in patients with a distinct ER the propagating wavefront

within the CTI was significantly narrower (24.92mm v 37.45mm, p=0.003) than in

those without a distinct ER.

Areas of conduction block with double potentials were found in 12 patients,

consistent with the Crista Terminalis (CT) as shown in Figure 2B. The leading edge

of the dominant conduction wavefront was defined accurately by high density

mapping and took a complex course around the right atrium, which was not

necessarily circumferential around the tricuspid annulus (Figure 3). In all patients the

conduction wavefront travelled both anteriorly and posteriorly to the SVC superior to

the line of conduction block produced by the crista. In 30% of patients with CCW

flutter, the posterior conduction wavefront reached the TA ahead of the anterior

9

wavefront and was thus deemed the dominant wavefront in the flutter circuit (Figure

3).

In contrast to the widely accepted belief that the CTI is the slowly conducting isthmus

of the circuit of typical AFL, our activation and propagation maps (Figure 4 and

Movie 1 in the Data Supplement) of the flutter circuit showed consistent CV around

the circuit with no particular slowing seen at the CTI. We analyzed this further by

assessing both gross CV across the lateral, CTI, septal and superior RA and local

CV at different points on the clock face. CV was not significantly different at the CTI

compared to the remainder of the RA (1.02 vs. 1.03 m/s, p = 0.93). When patients in

CW flutter were excluded, CV was still not significantly different at the CTI (1.06 vs.

1.06 m/s, p = 0.999). As shown in Figure 5A, in CCW flutter, no region of the RA had

a significantly different gross CV. Patients in CW also had no significant difference in

CV at the CTI (0.86 vs. 0.91 m/s, p = 0.81). Manual inspection of the activation maps

revealed areas of slow conduction in each CW flutter patient, however these areas

were not common between patients. There was no significant difference in regional

conduction velocity in CW vs. CCW flutter (Figure 2 Data Supplement)

Further analysis was performed excluding CW flutter patients. Local CV was

calculated using the distance travelled in 5ms at the centre of 8 equally spaced

points on the clock face. The average local CV was 1.22 m/s. Each activation map

showed areas of slow conduction, however these were heterogenous between

patients. When all 10 patients were averaged, there was no segment of the RA with

a significantly different local conduction velocity, this is shown in Figure 5B (one-way

ANOVA comparing all segments, p = 0.75).

10

The sawtooth pattern of the surface EKG in typical flutter was compared to the

position of the dominant wavefront in the circuit. The typical position of the activation

wavefront at the start of the downslope was at 9.9 ± 0.9 o’clock with 8 (of 10)

subjects starting at between 10 and 9 o’clock. At the end of the downslope, the

wavefront’s position was between 2 and 12 o’clock in all patients, 12.55 ± 0.6 o’clock

on average. Figure 3 in the data supplement shows a representative example of this.

Figure 6 shows the duration, distance travelled and conduction velocity in both

phases of the EKG. The downslope represented on average 73% ± 9% of the total

flutter cycle length (Fig 6A, downslope, 182 ± 21ms vs. upslope, 68 ± 29ms,

p<0.0001). There was no significant difference in conduction velocity during the

downslope compared to upslope (Fig 6B, 0.88 vs. 0.91 m/s, p=0.79). Therefore,

measurement of the distance covered by the wavefront during the downslope

showed a significantly longer distance covered than during the upslope (Fig 6C,

downslope, 157 ± 38 mm vs. upslope, 60 ± 24 mm, p=0.0004). In addition there was

a good correlation between distance covered by the activation wavefront and time of

the upslope/downslope (Fig 6D) confirming that distance covered and not conduction

velocity was the major factor changing between longer and shorter phases of the

sawtooth flutter wave.

11

Discussion

The flutter circuit has long been thought to be contingent on slow conduction in the

CTI region1. However, using recently available very high-resolution contact maps of

the flutter circuit we have found that the CV of the dominant activation wavefront, at

the CTI region is the same as in the remaining RA. These high-resolution maps have

allowed accurate identification of the path of the dominant flutter wavefront.

Correlation with the typical EKG pattern has shown slow conduction is not the cause

of the sawtooth appearance of the P wave in the EKG of typical flutter.

Conduction velocity of the flutter circuit

Previous lower resolution mapping studies have shown slower conduction in the

region of the CTI. Olshanky et al described in 10 patients the presence of an area of

slow conduction in the low right atrium2. They used a quadripolar catheter

sequentially placed in approximately 30 sites in the RA. Feld et al, showed that

patients with AFL had significantly slowed CTI conduction velocities compared to

patients free of AFL13, they did this by pacing various sites in sinus rhythm and

calculating gross conduction velocity. Hassankhani et al used a 3D mapping system,

CARTO, to further test this hypothesis and their results agreed with the consensus

demonstrating that conduction was particularly slowed in the medial isthmus and

inferior septal areas3. They divided the TVA into 8 segments, measuring conduction

velocity within 0.5-1cm of the annulus, with a minimum of 8 points mapped around

the TVA. Sawa et al used a similar method with similar findings, using CARTO

mapping and to pick two points on a line parallel to the TA, whilst also within 15mm

of the TA4, the two points used were between 5 and 20mm of each other.

12

Kinder et al used pacing and recording sites across the CTI to assess conduction

across the CTI, they did not find evidence of functional slowing or conduction delay

at faster cycle lengths or during AFL 14.

Studies using non-contact mapping have found the CTI region to have a significantly

slower CV 5,15. Schilling et al used non-contact mapping to investigate 13 patients

with AFL. They used short straight-line distances (<5mm) in the direction parallel to

wavefront propagation to calculate CV. They found that the CTI was significantly

slower than the remaining RA, however once patients with previous ablation were

removed, this difference was no longer statistically significant 16. As ablation in the

CTI region that had not fully abolished the atrial flutter circuit would be expected to

slow conduction in that area, the data excluding these patients are likely to be more

representative. Dixit et al also studied the flutter circuit using non contact mapping

and similarly found no differences in regional conduction time17.

In contrast to previous methods, high-resolution mapping has permitted accurate

identification of the dominant wavefront and leading edge at any part of the circuit. It

is possible that previous, lower resolution, studies identified regions of slow

conduction that were not part of the main flutter circuit. Previous studies have

artificially picked points close to the TA, with a wavefront direction parallel to the TA

3,4. With high-resolution mapping we were able to accurately determine the location

of the dominant wavefront, which was not always close to the TVA. Similarly,

conduction was not always parallel to the TVA. Therefore our data represents the CV

of the main flutter wavefront, whereas previous studies may have analyzed passive

13

areas that were not participating in the dominant wavefront. Although our previous

studies using non-contact mapping have also shown that wavefront distance from

the TVA varies around the flutter circuit 16, the resolution of this technique was

insufficient to define the leading edge of the conduction wavefront as accurately as in

this study. Looking at the CTI region, we have found that conduction velocity of the

wavefront of atrial flutter is not significantly different from the remainder of the RA.

We have shown that an understanding of the location of the leading edge of the

activation wavefront is vital in obtaining an accurate CV at the CTI because of

conduction slowing seen at the ER in the posterior aspect of the CTI. Sampling in

this region would give the false impression of conduction slowing at the CTI.

Explanation of the classic ‘Sawtooth’ EKG pattern

Although our study accurately defines cardiac activation during the different phases

of the sawtooth flutter wave and excludes slow conduction at the CTI as the cause

for the slow downstroke in the flutter EKG it cannot explain why the sawtooth pattern

occurs in the first place. The typical flutter waveform is composed of a gradual

downslope (also known as plateau or flat phase), sharp negative component, and

upstroke (which often overshoots)1, this is depicted in Figure 1 as the red, orange

and green lines respectively. A comparison of the vector of activation during the

upslope with the surface EKG vector of lead III (Data Supp. Figure 3) shows some

correlation but this is imprecise and certainly does not explain why within a small

segment of the TA circumference (approx. 1 o’clock to 10 o’clock) the EKG vector

reverses. In fact this would be impossible if the RA was activating in isolation. The

14

explanation must therefore in part relate to either activation of the LA or passive

activation of portions of the right atrium, in particular the posterior wall.

Indeed experimental studies have suggested the LA may be responsible for the

majority of the surface EKG pattern 7,18,19. Rosen et al demonstrated that rapid pacing

at the CS reproduces a sawtooth EKG similar to the typical atrial flutter 19. The typical

flutter sawtooth surface EKG pattern contains a sharp terminal negative deflection.

Using body surface mapping, Sippens, Groenwegen et al concluded that this steep

component of the downstroke corresponds to caudocranial activation of the

interatrial septum, while the medial portion of the subeustachian isthmus generated a

period of relative electrical silence 20. Bernstein et al showed that during termination

of AFL by ablation there was no sharp terminal negative component to the

downstroke, implying therefore that this component was a result of conduction past

the CTI region, into the interatrial septum and LA21. The authors concluded that the

sharp terminal negative deflection was unlikely to be solely a result of the

depolarization of the interatrial septum, given its relatively small mass in relation to

the rest of the atria. Adding further weight to this is evidence that patients’ with

previous LA ablation may have atypical surface EKG patterns when in typical CTI

dependent atrial flutter22, similarly LA enlargement or disease has been correlated

with terminal positivity of the flutter wave23. Our data showed that the terminal portion

of the downslope correlated with conduction around the interatrial septum thus

providing some support to the theory that either the interatrial septum or LA is

responsible for the sharp negative deflection of the sawtooth. Left atrial activation

therefore likely plays a major role in the genesis of the classic ‘sawtooth’ EKG

pattern.

15

In broad agreement with our findings are data from Sasaki et al24. They described

four phases to the flutter wave and found that the upslope indicated conduction on

the RA Free Wall from 12 to 8 o’clock similar to our findings showing this correlated

with the wavefront between1 to 10 o’clock.

Posterior boundaries of the flutter circuit

The Crista Terminalis (CT) is generally accepted as a major component of the

posterior border of the typical flutter circuit. The use of entrainment mapping and

intracardiac echocardiography (ICE) demonstrated that areas anterior to the CT are

within the re-entrant circuit, whereas areas just posterior to it are not 25. In our study,

12 out of 13 patients had clear evidence of a CT, which appeared to be the posterior

boundary of the flutter circuit. The ER has been described as an additional posterior

boundary to the flutter circuit between the coronary sinus ostium and the inferior

vena cava 25. Nakagawa et al described the ER as a region of fixed anatomical block

26. In our study we identified 50% of typical flutter patients with a manifest ER,

potentially reducing the length of isthmus ablation required to block the flutter circuit

at the CTI. However a previous attempt at this had mixed results27, with ablation at

this site having an increased risk of AV nodal damage. in some centers, it is

possible that through more detailed mapping with the Rhythmia mapping system,

these limitations could be overcome.

16

Limitations

As all patients were in AFL of obvious RA origin there was no clinical indication to

access the left atrium. We were therefore unable to perform LA maps, which may

have allowed further investigation into the contribution of LA activation to the surface

EKG in typical flutter. The broad definition of the sawtooth wave into upslope and

downslope was made because more nuanced distinction between slow (also

referred to as flat or plateau phase) and fast downslope, and brief overshoot of

upstroke could not be reliably discerned in all patients. While the manual

identification of start and end of upslope and downslope may not always be precise,

the variation due to this is unlikely to exceed a few ms and is therefore unlikely to

significantly impact on the results. Due to the novel nature of the Rhythmia mapping

system, only a small number of patients were studied. The point density of the 3D

maps varied from case to case due to operator and patient anatomy. Even the

lowest point density (52 points/cm2) is substantially more than achieved by previous

studies of the right atrium in atrial flutter.

17

Conclusion

Investigation of the circuit of typical CTI dependent atrial flutter with high-density

mapping has shown that the activation wavefront in AFL has a complex path around

the TA contributed to by the functional anatomy of the RA, in particular the presence

of the Crista Terminalis and Eustachian Ridge. Within this complex path, the

conduction velocity is not significantly slower at the CTI than in other regions and

therefore the gradual downslope of the sawtooth EKG is not due to slow conduction

at the CTI. This understanding was enabled by the use of the Rhythmia mapping

system producing accurate very high-density 3D maps which helped us to better

define the direction and velocity of the leading edge of the conduction wavefront in

different parts of the right atrium and in the context of the functional anatomy of the

chamber which was also well defined using this system. The knowledge that the

established ablation target of the CTI is not associated with slower conduction also

reinforces the message that ablation of an anatomical isthmus is often important in

terminating macroreentrant tachycardias.

Acknowledgements

We would like to acknowledge the BRC, BHF and ElectroCardioMaths Programme

of the Imperial Centre for Cardiac Engineering

18

References

1. Bun S-S, Latcu DG, Marchlinski F, Saoudi N. Atrial flutter: more than just one

of a kind. Eur Heart J. 2015;36:2356–2363.

2. Olshansky B, Okumura K, Hess PG, Waldo AL. Demonstration of an area of

slow conduction in human atrial flutter. J Am Coll Cardiol. 1990;16:1639–1648.

3. Hassankhani A, Yao B, Feld GK. Conduction velocity around the tricuspid

valve annulus during type 1 atrial flutter: defining the location of areas of slow

conduction by three-dimensional electroanatomical mapping. J Interv Card

Electrophysiol. 2003;8:121–127.

4. Sawa A, Shimizu A, Ueyama T, Yoshiga Y, Suzuki S, Sugi N, Matsuaki M.

Conduction Velocity around the Tricuspid Valve Annulus during Typical Atrial

Flutter by Electro-anatomic Mapping System. Journal of Arrhythmia.

2006;22:31–36.

5. Kondo M, Fukuda K, Wakayama Y, Nakano M, Hasebe Y, SHIMOKAWA H.

Usefulness of the Noncontact Mapping System to Elucidate the Conduction

Property for the Treatment of Common Atrial Flutter. Pacing and Clin

Electrophysiol. 2012;35:1464–1471.

6. Chen J, Hoff PI, Erga KS, Rossvoll O, Ohm O-J. Three-Dimensional

Noncontact Mapping Defines Two Zones of Slow Conduction in the Circuit of

Typical Atrial Flutter. Pacing and Clin Electrophysiol. 2003;26:318–322.

7. Okumura K, Plumb VJ, Pagé PL, Waldo AL. Atrial activation sequence during

atrial flutter in the canine pericarditis model and its effects on the polarity of the

19

flutter wave in the electrocardiogram. J Am Coll Cardiol. 1991;17:509–518.

8. Rottner L, Metzner A, Ouyang F, Heeger C, Hayashi K, Fink T, Lemes C,

Mathew S, Maurer T, Reißmann B, Rexha E, Riedl J, Saguner AM, Santoro F,

Kuck K-H, Sohns C. Direct Comparison of Point-by-Point and Rapid Ultra-

High-Resolution Electroanatomical Mapping in Patients Scheduled for Ablation

of Atrial Fibrillation. J Cardiovasc Electrophysiol. 2017;28:289–297.

9. Schaeffer B, Hoffmann BA, Meyer C, Akbulak RÖ, Moser J, Jularic M, Eickholt

C, Nührich JM, Kuklik P, Willems S. Characterization, Mapping, and Ablation

of Complex Atrial Tachycardia: Initial Experience With a Novel Method of Ultra

High-Density 3D Mapping. J Cardiovasc Electrophysiol. 2016;27:1139–1150.

10. Nakagawa H, Ikeda A, Sharma T, Lazzara R, Jackman WM. Rapid High

Resolution Electroanatomical Mapping: Evaluation of a New System in a

Canine Atrial Linear Lesion Model. Circ Arrhythm Electrophysiol. 2012;5:417–

424.

11. Anter E, Tschabrunn CM, Contreras-Valdes FM, Li J, Josephson ME.

Pulmonary vein isolation using the Rhythmia mapping system: Verification of

intracardiac signals using the Orion mini-basket catheter. Heart Rhythm.

2015;12:1–8.

12. Bollmann A, Hilbert S, John S, Kosiuk J, Hindricks G. Insights from preclinical

ultra high-density electroanatomical sinus node mapping. Europace.

2015;17:489–494.

13. Feld GK, Mollerus M, Birgersdotter-Green U, Fujimura O, Bahnson TD, Boyce

K, Rahme M. Conduction Velocity in the Tricuspid Valve-Inferior Vena Cava

20

Isthmus is Slower in Patients With Type I Atrial Flutter Compared to Those

Without a History of Atrial Flutter. J Cardiovasc Electrophysiol. 1997;8:1338–

1348.

14. Kinder C, Kall J, Kopp D, Rubenstein D, Burke M, Wilber D. Conduction

properties of the inferior vena cava tricuspid annular isthmus in patients with

typical atrial flutter. J Cardiovasc Electrophysiol. 1997;8:727–737.

15. Chen J, Hoff PW, Erga KS, Rossvoll O, Ohm OJ. Three-dimensional

noncontact mapping defines two zones of slow conduction in the circuit of

typical atrial flutter. Pacing and Clin Electrophysiol. 2003;26:318–322.

16. Schilling RJ, Peters NS, Goldberger J, Kadish AH, Davies DW.

Characterization of the anatomy and conduction velocities of the human right

atrial flutter circuit determined by noncontact mapping. J Am Coll Cardiol.

2001;38:385–393.

17. Dixit S, Lavi N, Robinson M, Riley MP, Callans DJ, Marchlinski FE, Lin D.

Noncontact Electroanatomic Mapping to Characterize Typical Atrial Flutter:

Participation of Right Atrial Posterior Wall in the Reentrant Circuit. J

Cardiovasc Electrophysiol. 2010;22:422–430.

18. Schoels W, Offner B, Brachmann J, Kuebler W, Elsherif N. Circus Movement

Atrial-Flutter in the Canine Sterile Pericarditis Model - Relation of

Characteristics of the Surface Electrocardiogram and Conduction Properties of

the Reentrant Pathway. J Am Coll Cardiol. 1994;23:799–808.

19. Rosen KM, Lau SH, Damato AN. Simulation of atrial flutter by rapid coronary

sinus pacing. Am Heart J 1969;78:635–642.

21

20. SippensGroenewegen A, Lesh MD, Roithinger FX, Ellis WS, Steiner PR,

Saxon LA, Lee RJ, Scheinman MM. Body surface mapping of

counterclockwise and clockwise typical atrial flutter: a comparative analysis

with endocardial activation sequence mapping. J Am Coll Cardiol.

2000;35:1276–1287.

21. Bernstein NE, Sandler DA, Goh M, Feigenblum DY, Holmes DS, Chinitz LA.

Why a Sawtooth? Inferences on the Generation of the Flutter Wave during

Typical Atrial Flutter Drawn from Radiofrequency Ablation. Annals of

Noninvasive Electrocardiology. 2004;9:358–361.

22. Chugh A. Characteristics of Cavotricuspid Isthmus-Dependent Atrial Flutter

After Left Atrial Ablation of Atrial Fibrillation. Circulation. 2006;113:609–615.

23. Milliez P, Richardson AW, Obioha-Ngwu O, Zimetbaum PJ, Papageorgiou P,

Josephson ME. Variable electrocardiographic characteristics of isthmus-

dependent atrial flutter. J Am Coll Cardiol. 2002;40:1125–1132.

24. Sasaki K, Sasaki S, Kimura M, Owada S, Horiuchi D, Itoh T, Ishida Y,

Okumura K. Revisit of Typical Counterclockwise Atrial Flutter Wave in the

ECG: Electroanatomic Studies on the Determinants of the Morphology. Pacing

and Clin Electrophysiol. 2013;36:978–987.

25. Olgin JE, Kalman JM, Fitzpatrick AP, Lesh MD. Role of Right Atrial

Endocardial Structures as Barriers to Conduction During Human Type I Atrial

Flutter : Activation and Entrainment Mapping Guided by Intracardiac

Echocardiography. Circulation. 1995;92:1839–1848.

26. Nakagawa H, Lazzara R, Khastgir T, Beckman KJ, McClelland JH, Imai S,

22

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 septal isthmus and a new technique for rapid

identification of ablation success. Circulation. 1996;94:407–424.

27. Anselme F, Klug D, Scanu P, Poty H, Lacroix D, Kacet S, Cribier A, Saoudi N.

Randomized comparison of two targets in typical atrial flutter ablation. Am J

Cardiol. 2000;85:1302–1307.

23

Tables

Table 1. Procedural data

Case No.

Sex Age Direction Cycle length (ms)

Points on map

Volume of RA (cm2)

Surface area of RA (cm2)

Point density (points/cm2)

Time to create flutter map (mm:ss)

Points collected (points/sec)

1 M 56 CCW 204 20989 147 86 243 11:16 31.052 F 75 CCW 274 23687 221 171 138 14:17 27.643 M 69 CCW 278 11091 184 138 82 23:14 8.094 M 69 CCW 249 13199 126 105 126 16:05 13.685 M 71 CCW 250 13074 160 131 100 23:08 9.426 M 61 CCW 331 7111 154 119 60 14:35 8.137 M 74 CCW 196 10283 163 129 79 12:41 13.518 M 81 CCW 240 46157 268 120 385 38:21 20.069 M 64 CCW 240 10883 141 124 88 14:01 12.94

10 M 68 CCW 236 18509 102 105 176 16:27 18.7511 M 67 CW 256 16710 209 177 95 16:44 16.6412 M 57 CW 255 19032 182 137 139 24:30 12.9513 M 65 CW 314 7256 209 141 52 21:37 5.59

Mean 67±7

255±38 16767±10233

174±44 129±25 136±91 19:00±07:16 15±8

CW: clockwise, CCW: counter clockwise, RA: right atrium

24

Figure 1

25

B

Figure 2

26

Figure 3

27

Figure 4

28

Figure 5

29

BA

Figure 6

Figure 5

The upstroke and downstroke of the sawtooth flutter wave compared to conduction

velocity and distance travelled by the wavefront, in patients in CCW flutter. A, The

downstroke phase of the sawtooth flutter waveform EKG is significantly longer in

duration than the upstroke. B, There is no difference in CV between upstroke and

downstroke of the EKG. C, The flutter wavefront in the downstroke phase of the EKG

covers a significantly longer distance. D, Distance travelled by the wavefront and

duration of each phase of the EKG is correlated

30

A B

C D

Figure legends

Figure 1

Surface EKG leads II, III and AVF showing a typical flutter waveform. This is

composed of a gradual downslope, sharp negative component, and upstroke shown

as the red, orange and green lines respectively.

Figure 2

The Eustachian ridge and Crista Terminalis in Atrial flutter. The window of interest

has been manually altered to half of the cycle length. Color wheel shown in top right.

A, Local activation time (LAT) map showing area of block and double potentials (*)

in the inferior segment of the CTI, consistent with the Eustachian ridge. B, LAT map

showing area of block and double potentials (*) at the lateral RA wall consistent with

the Crista Terminalis.

IVC: Inferior vena cava, SVC: Superior vena cava, TVA: Tricuspid valve annulus

Figure 3

Local activation time (LAT) maps showing complex activation of the flutter

wavefront. A, shows the septal wall with activation proceeding both around the TA

(red arrow) but more rapidly in a posterior direction across the free wall and over the

Crista Terminalis. B, The wavefront from the free wall enters the lateral TA region

(white arrow) ahead of the wavefront passing circumferentially around the TA (red

arrow). These images demonstrate the complex course of the flutter wavefront,

which does not necessarily take a circumferential path around the TA.

IVC: Inferior vena cava, SVC: Superior vena cava

31

Figure 4

Local activation time (LAT) map showing the inferior portion of the RA, including

the CTI. Each color isochrone represents 34.75ms. There is no evident region of

slow conduction in the region of the CTI.

IVC: Inferior vena cava, SVC: Superior vena cava, TVA: Tricuspid valve annulus

Figure 5

Conduction velocity around the RA in CCW AFL. A, CV at each area of the RA in

patients in CCW AFL, B, CV at each point as defined by placing the TVA in the LAO

position, in patients in CCW AFL

NS: Nonsignificant (p >0.05)

IVC: Inferior vena cava, SVC: Superior vena cava, TVA: Tricuspid valve annulus

Figure 6

The upstroke and downstroke of the sawtooth flutter wave compared to conduction

velocity and distance travelled by the wavefront, in patients in CCW flutter. A, The

downstroke phase of the sawtooth flutter waveform EKG is significantly longer in

duration than the upstroke. B, There is no difference in CV between upstroke and

downstroke of the EKG. C, The flutter wavefront in the downstroke phase of the EKG

covers a significantly longer distance. D, Distance travelled by the wavefront and

duration of each phase of the EKG is correlated

32