characterisation of the substrate of atrial fibrillation and flutter

Title Page

Characterisation of the Substrate of

Atrial Fibrillation and Flutter

Martin Kingsland Stiles

MB ChB, FRACP

Department of Cardiology

Royal Adelaide Hospital

and

Discipline of Medicine

University of Adelaide

A thesis submitted to the University of Adelaide in

fulfilment of the requirements for the degree of

Doctor of Philosophy

August 2008

II

Dedication

To my wife Christina and my children Isabelle and Liam

III

Table of Contents

Abstract .................................................................................. XIII

Declaration ..............................................................................XV

Acknowledgements ................................................................XVI

Publications and Communications to Learned Societies .........XVII

Prizes and Awards during Candidature....................................XXI

Abbreviations ........................................................................XXII

Chapter 1. Review of the Literature1

1.1 Introduction ................................................................................................... 1

1.1.1 History of Atrial Fibrillation................................................................................... 2

1.1.2 Epidemiology of Atrial Fibrillation ........................................................................ 3

1.1.3 History of Atrial Flutter ......................................................................................... 9

1.1.4 Epidemiology of Atrial Flutter............................................................................. 11

1.2 The Electrophysiological Basis of Atrial Fibrillation....................................... 13

1.2.1 The Multiple Heterotopous Centres Theory....................................................... 13

1.2.2 The Multiple Wavelet Hypothesis ...................................................................... 13

1.2.3 Localised Sources in the Initiation and Maintenance of Atrial Fibrillation ........ 15

1.2.4 The Venous Wave Hypothesis ............................................................................ 19

1.2.5 Rotors and Activation Gradients......................................................................... 21

IV

1.2.6 Summary ............................................................................................................. 22

1.3 Triggers and Substrate in Atrial Arrhythmia.................................................. 22

1.3.1 Focal Triggers ...................................................................................................... 23

1.3.2 Localised Reentry................................................................................................ 24

1.3.3 Anatomical Structures with Site‐specific Conduction Properties....................... 26

1.3.4 Complex Fractionated Atrial Electrograms......................................................... 29

1.3.5 Dominant Frequency .......................................................................................... 32

1.3.6 The Neural Basis of Atrial Fibrillation ................................................................. 34

1.3.7 Substrate Associated with Atrial Flutter............................................................. 36

1.3.8 Typical and Atypical Atrial Flutter....................................................................... 38

1.4 Inter‐relationships between Atrial Flutter and Atrial Fibrillation .................. 39

1.4.1 Atrial Flutter begets Atrial Fibrillation ................................................................ 40

1.4.2 Atrial Fibrillation begets Atrial Flutter ................................................................ 41

1.5 Rate‐related Remodelling............................................................................. 43

1.5.1 Electrical Remodelling ........................................................................................ 43

1.5.2 Structural Remodelling ....................................................................................... 45

1.6 Atrial Substrate in Conditions Associated with Atrial Arrhythmia ................. 47

1.6.1 Heart Failure ....................................................................................................... 47

1.6.2 Atrial Septal Defects ........................................................................................... 48

1.6.3 Hypertension....................................................................................................... 48

1.6.4 Valvular Heart Disease........................................................................................ 49

1.6.5 Sub‐clinical Atrial Disease in Lone Atrial Fibrillation .......................................... 51

1.7 Sinus Node Disease, Remodelling and Atrial Arrhythmias............................. 52

1.7.1 The Anatomical Sinus Node................................................................................ 52

V

1.7.2 The Functional Sinus Node Complex .................................................................. 53

1.7.3 Conventional Evaluation of Sinus Node Function .............................................. 56

1.7.4 High‐density Mapping of the Sinus Node ........................................................... 57

1.7.5 Sinus Node Disease............................................................................................. 58

1.7.6 Clinical Conditions Associated with Sinus Node Impairment............................. 59

Chapter 2. Right and left atrial electrical substrate of lone atrial fibrillation61

2.1 Introduction ................................................................................................. 61

2.2 Methods....................................................................................................... 62

2.2.1 Study Population................................................................................................. 62

2.2.2 Electrophysiological Study and Ablation ............................................................ 63

2.2.3 Electrophysiology Study Protocol ....................................................................... 64

2.2.4 Statistical Analysis............................................................................................... 66

2.3 Results.......................................................................................................... 67

2.3.1 Baseline Details................................................................................................... 67

2.3.2 Atrial Refractoriness ........................................................................................... 67

2.3.3 Atrial Conduction Time ....................................................................................... 68

2.3.4 Site‐Specific Conduction Abnormalities ............................................................. 68

2.3.5 Sinus Node Function ........................................................................................... 69

2.4 Discussion..................................................................................................... 69

2.4.1 Major findings ..................................................................................................... 69

2.4.2 Rate‐Related Electrical Remodelling................................................................... 70

2.4.3 Progression of Atrial Fibrillation ......................................................................... 71

2.4.4 Substrate Predisposing to the Development of Atrial Fibrillation ..................... 72

VI

2.4.5 Implications......................................................................................................... 74

2.4.6 Limitations .......................................................................................................... 74

2.5 Conclusion.................................................................................................... 75

Chapter 3. Right and left atrial electroanatomical substrate of lone atrial

fibrillation‐‐‐‐ ........................................................................... 82

3.1 Introduction ................................................................................................. 82

3.2 Methods....................................................................................................... 83

3.2.1 Study Population................................................................................................. 83

3.2.2 Electrophysiological Study and Ablation ............................................................ 84

3.2.3 Electroanatomical Study Protocol ...................................................................... 85

3.2.4 Statistical Analysis............................................................................................... 87

3.3 Results.......................................................................................................... 88

3.3.1 Baseline Details................................................................................................... 88

3.3.2 Structural and Voltage Abnormalities ................................................................ 88

3.3.3 Abnormalities in Conduction Velocity ................................................................ 89

3.3.4 Complex Electrograms ........................................................................................ 89

3.4 Discussion..................................................................................................... 90

3.4.1 Major findings ..................................................................................................... 90

3.4.2 Structural changes in Atrial Fibrillation .............................................................. 91

3.4.3 Electroanatomical Studies in Atrial Fibrillation and Related Conditions............ 91

3.4.4 Prior Studies Implicating Abnormalities in “Lone” Atrial Fibrillation ................. 93

3.4.5 Implications......................................................................................................... 93

VII

3.4.6 Limitations .......................................................................................................... 94

3.5 Conclusion.................................................................................................... 94

Chapter 4. Right atrial electrical substrate of atrial flutter ...................... 100

4.1 Introduction ................................................................................................100

4.2 Methods......................................................................................................101

4.2.1 Study Population............................................................................................... 101

4.2.2 Electrophysiological Study and Ablation .......................................................... 102

4.2.3 Electrophysiology Study Protocol ..................................................................... 102

4.2.4 Statistical Analysis............................................................................................. 104

4.3 Results.........................................................................................................105

4.3.1 Baseline Details................................................................................................. 105

4.3.2 Atrial Refractoriness ......................................................................................... 105

4.3.3 Atrial Conduction Time ..................................................................................... 106

4.3.4 Site‐Specific Conduction Abnormalities ........................................................... 106

4.3.5 Sinus Node Function ......................................................................................... 107

4.4 Discussion....................................................................................................107

4.4.1 Major Findings .................................................................................................. 107

4.4.2 The Inter‐relationship between Atrial Flutter and Atrial Fibrillation ............... 108

4.4.3 Substrate Predisposing to the Development of Atrial Flutter.......................... 109

4.4.4 Substrate in Conditions Predisposed to Atrial Arrhythmia .............................. 110

4.4.5 Implications....................................................................................................... 112

4.4.6 Limitations ........................................................................................................ 112

4.5 Conclusion...................................................................................................113

VIII

Chapter 5. Right atrial electroanatomical substrate of atrial flutter ........ 118

5.1 Introduction ................................................................................................118

5.2 Methods......................................................................................................119



5.2.3 Electroanatomical Study Protocol .................................................................... 120


5.3 Results.........................................................................................................123


5.3.2 Structural and Voltage Abnormalities .............................................................. 123

5.3.3 Abnormalities in Conduction Velocity .............................................................. 124

5.3.4 Complex Electrograms ...................................................................................... 124

5.4 Discussion....................................................................................................125

5.4.1 Major Findings .................................................................................................. 125

5.4.2 Substrate in Clinical Conditions Associated with Atrial Arrhythmia................. 125

5.4.3 Substrate in Experimental Models of Atrial Flutter.......................................... 126

5.4.4 Substrate in Clinical Atrial Flutter ..................................................................... 127

5.4.5 Implications....................................................................................................... 128

5.4.6 Limitations ........................................................................................................ 128

5.5 Conclusion...................................................................................................129

IX

Chapter 6. High‐density mapping of the sinus node to characterise the

remodelling resulting from chronic atrial flutter .................... 134

6.1 Introduction ................................................................................................134

6.2 Methods......................................................................................................135



6.2.3 Conventional Evaluation of Sinus Node Function ............................................ 137

6.2.4 High‐Density Simultaneous Unipolar Mapping of Sinus Node Function.......... 138


6.3 Results.........................................................................................................141


6.3.2 Sinus Node Dysfunction and Atrial Remodelling .............................................. 141

6.3.3 Extent and Shift of the Atrial Pacemaker Complex .......................................... 142

6.3.4 Preferential Pathway Conduction..................................................................... 143

6.3.5 Beat‐to‐beat Variation in Sinus Node Function................................................ 143

6.3.6 Crista Terminalis Conduction............................................................................ 144

6.3.7 Voltage Findings................................................................................................ 144

6.4 Discussion....................................................................................................145

6.4.1 Major Findings .................................................................................................. 145

6.4.2 The Sinus Node: Anatomical and Functional Considerations........................... 146

6.4.3 Atrial Remodelling and Sinus Node Function ................................................... 147

6.4.4 Preferential Pathways of Conduction............................................................... 149

6.4.5 Implications....................................................................................................... 151

X

6.4.6 Limitations ........................................................................................................ 151

6.5 Conclusion...................................................................................................152

Chapter 7. High‐density mapping of atrial fibrillation to determine the

optimal recording duration for dominant frequency and

automated detection of complex fractionated electrograms.. 166

7.1 Introduction ................................................................................................166

7.2 Methods......................................................................................................167


7.2.2 Electrophysiological Study ................................................................................ 168

7.2.3 High‐Density Bi‐atrial Mapping......................................................................... 168

7.2.4 Complex Fractionated Atrial Electrograms....................................................... 169

7.2.5 Dominant Frequency ........................................................................................ 170


7.3 Results.........................................................................................................171

7.3.1 Baseline details ................................................................................................. 171

7.3.2 CFE‐mean .......................................................................................................... 172


7.4 Discussion....................................................................................................173

7.4.1 Major Findings .................................................................................................. 173

7.4.2 Time Domain Analysis....................................................................................... 175

7.4.3 Frequency Domain Analysis.............................................................................. 176

7.4.4 Differences in Effect of Electrogram Duration between CFE‐mean and DF..... 177

XI

7.4.5 Limitations ........................................................................................................ 178

7.5 Conclusion...................................................................................................178


relationship between activation frequency, complex

fractionated electrograms and the anatomical substrate....... 187

8.1 Introduction ................................................................................................187

8.2 Methods......................................................................................................188



8.2.3 Study Protocol................................................................................................... 190


8.2.5 Activation Frequency ........................................................................................ 192

8.2.6 Relationship between Activation Frequency and Fractionation ...................... 192


8.3 Results.........................................................................................................194




8.3.4 Relationship between CFE‐mean and Dominant Frequency............................ 196

8.4 Discussion....................................................................................................198

8.4.1 Major findings ................................................................................................... 198

8.4.2 Spatial Distribution of Complex Fractionated Atrial Electrograms .................. 199

XII

8.4.3 Spatial Distribution of Dominant Frequency .................................................... 199

8.4.4 Relationship between Activation Frequency and Fractionation ...................... 201

8.4.5 Clinical Implications .......................................................................................... 202

8.4.6 Limitations ........................................................................................................ 203

8.5 Conclusion...................................................................................................203

Chapter 9. Summary‐‐‐‐‐.......................................................................... 213

Chapter 10. Future Directions‐‐‐ ................................................................ 220

References ........................................................................... 223

XIII

Abstract

Atrial fibrillation and atrial flutter are the most common sustained arrhythmias,

however their underlying mechanisms are yet to be fully characterised. This thesis

evaluates the electrophysiological and electroanatomical substrate of the atria in

patients with these arrhythmias.

Experimental studies of atrial fibrillation have demonstrated effective refractory

period shortening and conduction slowing as a result of atrial fibrillation giving rise to

the concept that “atrial fibrillation begets atrial fibrillation”. However, cardioversion to

prevent electrical remodelling does not prevent progression of disease, suggesting a

“second factor” drives this process. Chapters 2 and 3 evaluate the atrial substrate in

patients with “lone” atrial fibrillation. These studies demonstrate such patients,

remote from an arrhythmic event, have prolongation of atrial refractoriness,

conduction slowing, impairment of sinus node function, site‐specific conduction delay,

lower voltage and a greater proportion of complex electrograms compared to

reference patients. These abnormalities constitute the “second factor” critical to the

development and progression of atrial fibrillation.

Atrial flutter has a close inter‐relationship with atrial fibrillation and these rhythms

frequently co‐exist. Atrial fibrillation often occurs in patients with heart disease known

to demonstrate abnormal atrial substrate; whether similar substrate exists in patients

with atrial flutter to account for the co‐existence of both arrhythmias is unknown.

Chapters 4 and 5 evaluate the atrial substrate in patients with atrial flutter, remote

from arrhythmia, demonstrating structural abnormalities characterised by loss of

myocardial voltage, conduction slowing and impaired sinus node function, without

XIV

reduction in atrial refractoriness. These findings implicate a common substrate as the

cause of the close inter‐relationship between these arrhythmias.

There is a frequent association between atrial arrhythmia and sinus node disease for

which several mechanisms have been postulated. In addition, there is a size

discrepancy between the anatomical sinus node and the much larger functional sinus

node complex. Little is known about normal sinus node function or the effects of

remodelling due to arrhythmia. Chapter 6 characterises sinus node activation to

determine the nature and extent of the functional sinus node complex in patients with

and without chronic atrial flutter. The functional sinus node complex demonstrates

dynamic shifts in activation with preferential pathways of conduction to atrial

myocardium. Patients with atrial flutter demonstrate lesser voltage, longer conduction

times along preferential pathways and a smaller functional sinus node complex. These

findings provide insights into the function of the human sinus node in health and

disease.

Sites of complex fractionated atrial electrograms and highest dominant frequency are

implicated in maintaining atrial fibrillation. Chapter 7 determines the minimum

recording duration that accurately characterises electrogram complexity and activation

frequency. An electrogram duration of ≥5 seconds is required to accurately identify

these sites. Chapter 8 evaluates the relationship between sites of fractionation and

high frequency activation during atrial fibrillation. Greater fractionation and higher

dominant frequency are seen in persistent atrial fibrillation and left atria. Preferential

areas of high dominant frequency are observed in paroxysmal but not persistent atrial

fibrillation. Areas of complex fractionated atrial electrograms are found adjacent to

sites of high dominant frequency.

XV

Declaration

This work contains no material which has been accepted for the award of any other

degree or diploma in any university or other tertiary institution to Martin Stiles and, to

the best of my knowledge and belief, contains no material previously published or

written by another person, except where due reference has been made in the text.

I give consent to this copy of my thesis when deposited in the University Library, being

made available for loan and photocopying, subject to the provisions of the Copyright

Act 1968.

The author acknowledges that copyright of published works contained within this

thesis (as listed below) resides with the copyright holder(s) of those works.

Martin Kingsland Stiles ___________________________________________ August 2008

Stiles MK, John B, Wong CX, Kuklik P, Brooks AG, Lau DH, Dimitri H, Roberts‐Thomson KC, Wilson L, De Sciscio P, Young GD, Sanders P. Paroxysmal lone atrial fibrillation is associated with abnormal atrial substrate: Characterising the “second factor”. J Am Coll Cardiol, in press

Stiles MK, Brooks AG, Kuklik P, John B, Dimitri H, Lau DH, Wilson L, Shashidhar, Roberts‐Thomson RL, Mackenzie L, Young GD, Sanders P. High‐density mapping of atrial fibrillation in humans: Relationship between high frequency activation and electrogram fractionation. J Cardiovasc Electrophysiol, 2008;19:1245‐1253

Stiles MK, Brooks AG, Sanders P. Putting CFAE on the map. J Cardiovasc Electrophysiol 2008;19:904‐906

Stiles MK, Brooks AG, John B, Shashidhar, Wilson L, Kuklik P, Dimitri H, Lau DH, Roberts‐Thomson RL, Mackenzie L, Willoughby S, Young GD, Sanders P. Effect of electrogram duration on the quantification of complex fractionated atrial electrograms and dominant frequency. J Cardiovasc Electrophysiol 2008;19:252‐258

Sanders P, Stiles MK, Young GD. Virtual Anatomy for Atrial Fibrillation Ablation. J Cardiovasc Electrophysiol 2006;17:349‐351

XVI

Acknowledgements

I would like to thank Professor Prashanthan Sanders and Dr Glenn Young as my supervisors and mentors. It has been a privilege and a pleasure to work with them and to learn from them and I admire their dedication to and knowledge of their chosen field. I am certain our collaboration and friendship will continue well into the future. I am grateful to the New Zealand Heart Foundation for the Overseas Training Fellowship which financially supported my first year of study, and the Royal Adelaide Hospital for the Dawes Scholarship which supported me subsequently. I am also indebted to the patients who, despite the anxieties associated with medical intervention, volunteered additional time during their procedure for me to gather the data on which this thesis is based. I am thankful to the Fellows who made the Electrophysiology Department a delight to work in, especially to Drs Bobby John, Dennis Lau, Hany Dimitri, Shashidhar, Han Lim, Muayad Alasady, Gautam Sharma and Narayanan Namboodiri. Special thanks goes to Mr Christopher Wong whose many hours of work performing data analysis kept my project on time, and to Dr Anthony Brooks whose hard work and good humour assisted me to produce this work. I am grateful to Dr Pawel Kuklik who developed the software critical to my data analysis and to the other members of the Cardiovascular Research Centre who provided the essential academic environment to nourish critical thought and independent analysis. Additional thanks to Mr Thomas Sullivan for statistical analysis. I am indebted to Professor Jonathan Kalman whose expertise in sinus node function he shared to significantly improve Chapter 6. My sincere appreciation goes to Dr Graydon Beatty for his help in understanding the complexities of the non‐contact mapping system and the mysteries of the unipolar electrogram. I acknowledge the encouragement of Associate Professor Warren Smith to follow an academic path in Adelaide. I would also like to thank the staff of the Cardiovascular Investigation Unit, Royal Adelaide Hospital without whom this research could not have taken place, in particular Ms Lauren Wilson for her encouragement and friendship. A special thanks to my parents Roger and Joan whose support, encouragement and love have always been there for me. It is because of them that our family has been empowered to seek their own course in life and, no matter where in the world this took us, the knowledge of their unqualified support was always there to sustain us. Most importantly, I wish to thank Christina, Isabelle and Liam whom I love dearly and to whom this thesis is dedicated. Christina has travelled with me, literally and figuratively, throughout my journey of learning whilst remaining understanding and supportive of the additional demands this has made on us all. I am sure we will look back fondly on our time in Adelaide and the friends we made here.

XVII

Publications and Communications to Learned Societies

Chapter 1

i) Manuscript: Stiles MK, Brooks AG, Sanders P. Putting CFAE on the map. J

Cardiovasc Electrophysiol 2008;19:904‐906

ii) Manuscript: Sanders P, Stiles MK, Young GD. Virtual Anatomy for Atrial

Fibrillation Ablation. J Cardiovasc Electrophysiol 2006;17:349‐351

iii) Presentation: Invited to present at the Cardiac Society of Australia and New

Zealand 56th Scientific Meeting August 2008, Adelaide, Australia on

Complex Electrograms and Dominant Frequency

Chapters 2 and 3

i) Manuscript: Stiles MK, John B, Wong CX, Kuklik P, Brooks AG, Lau DH,

Dimitri H, Roberts‐Thomson KC, Wilson L, De Sciscio P, Young GD, Sanders

P. Paroxysmal lone atrial fibrillation is associated with abnormal atrial

substrate: Characterising the “second factor”. J Am Coll Cardiol, in press

ii) Presentation: Presented at the Cardiac Society of Australia and New

Zealand 56th Scientific Meeting August 2008, Adelaide, Australia and

published in abstract form (Heart Lung Circ 2008;17:S120)

iii) Presentation: Presented at the Medical Grand Round, Royal Adelaide

Hospital, August 2008. Finalist for the 2008 Nimmo Prize for Full‐Time

Research

XVIII

Chapters 4 and 5

i) Manuscript: Stiles MK, Wong CX, John B, Kuklik P, Brooks AG, Lau DH,

Dimitri H, Wilson L, Young GD, Sanders P. Atrial flutter is associated with

diffuse atrial abnormalities: Implications for the development of atrial

fibrillation. Submitted for publication

ii) Presentation: Presented at the Cardiac Society of Australia and New



Chapter 6

i) Manuscript: Stiles MK, Brooks AG, Roberts‐Thomson KC, Kuklik P, John B,

Young GD, Kalman JM, Sanders P. High‐density mapping of the sinus node

in humans: role of preferential pathways and the effect of remodelling.

Submitted for publication

ii) Presentation: Presented at the Heart Rhythm Society 28th Annual Scientific

Sessions, May 2008, San Francisco, United States of America and published

in abstract form (Heart Rhythm 2008;5:S278)

iii) Presentation: Presented at the Cardiac Society of Australia and New



XIX

Chapter 7

i) Manuscript: Stiles MK, Brooks AG, John B, Shashidhar, Wilson L, Kuklik P,

Dimitri H, Lau DH, Roberts‐Thomson RL, Mackenzie L, Willoughby S, Young

GD, Sanders P. Effect of electrogram duration on the quantification of

complex fractionated atrial electrograms and dominant frequency. J

Cardiovasc Electrophysiol 2008;19:252‐258

ii) Presentation: Presented at the Heart Rhythm Society 28th Annual Scientific

Sessions, May 2007, Denver, United States of America and published in

abstract form (Heart Rhythm 2007;4:S348)


Zealand 55th Scientific Meeting August 2007, Christchurch, New Zealand

and published in abstract form (Heart Lung Circ 2007;16:S109)

Chapter 8

i) Manuscript: Stiles MK, Brooks AG, Kuklik P, John B, Dimitri H, Lau DH,

Wilson L, Shashidhar, Roberts‐Thomson RL, Mackenzie L, Young GD,

Sanders P. High‐density mapping of atrial fibrillation in humans:

Relationship between high frequency activation and electrogram

fractionation. J Cardiovasc Electrophysiol 2008;19:1245‐1253

ii) Presentation: Presented at the American College of Cardiology 57th Annual

Scientific Sessions, March 2008, Chicago, United States of America and

published in abstract form (J Am Coll Cardiol 2008;51:A6)

XX




iv) Presentation: Presented at the Post‐graduate Research Expo, School of

Medicine, July 2008

XXI

Prizes and Awards during Candidature

i) New Zealand Heart Foundation Overseas Training Fellowship, 2005‐2006

ii) New Zealand Heart Foundation Travel Grant, Heart Rhythm 06, Boston,

United States of America, 2006

iii) Dawes Scholarship, Royal Adelaide Hospital, 2006‐2008

iv) Cardiac Society of Australia and New Zealand Annual Scientific Meeting

Travelling Scholarship, Canberra, 2006

v) St Jude Medical Australia Pacing and Electrophysiology Young Investigator

Award, 2006

vi) Cardiac Society of Australia and New Zealand Travelling Scholarship,

European Society of Cardiology Congress, Barcelona, Spain, 2006

vii) Best Moderated Poster in the Drugs/Ablation/Electrical Therapy of

Arrhythmias section, World Congress of Cardiology, Barcelona, Spain, 2006

viii) Australian Heart Foundation Travel Grant, Heart Rhythm 07, Denver, United

States of America, 2007

ix) University of Adelaide, Postgraduate Travelling Fellowship, 2008

x) Australian Heart Foundation Travel Grant, Heart Rhythm 08, San Francisco,

United States of America, 2008

xi) Finalist for the Nimmo Prize for Full‐Time Research, Adelaide, 2008

XXII

Abbreviations

CFAE Complex Fractionated Atrial Electrogram

CFE‐mean An automated measure of electrogram complexity

CI Confidence Interval

CSNRT Corrected Sinus Node Recovery Time

DF Dominant Frequency

EA Earliest Activation

ECG Electrocardiogram

ERP Effective Refractory Period

ICC Intra‐Class Correlation

IQR Inter‐Quartile Range

OR Odds Ratio

SACT Sino‐Atrial Conduction Time

SBO Sinus Break‐Out

Standard international NBG pacing codes are used (e.g. VVI, DDD)

Due to the common initials of atrial flutter and atrial fibrillation, these rhythms are

not abbreviated in the text. For some tables and figures, the abbreviation “AF” is

used for atrial fibrillation. Atrial flutter is abbreviated to “flutter” in some tables

and figures.

1

Chapter 1. Review of the Literature‐__

1.1 Introduction

Atrial fibrillation is a supraventricular tachyarrhythmia characterized by uncoordinated

atrial activation with consequent deterioration of atrial mechanical function. On the

electrocardiogram, atrial fibrillation is characterized by the replacement of consistent P

waves by rapid oscillations or fibrillatory waves that vary in amplitude, shape and

timing, associated with an irregular and often rapid ventricular response when

atrioventricular conduction is intact. The ventricular response to atrial fibrillation

depends on electrophysiological properties of the atrioventricular node and other

conducting tissues, the level of vagal and sympathetic tone, and the action of drugs.

ACC/AHA/ESC 2006 Guidelines for the Management of Patients with Atrial Fibrillation

(Fuster et al. 2006).

Atrial flutter is characterized by an organized atrial rhythm with a rate typically

between 250 and 350 bpm. Electrophysiological studies have shown that this simple

ECG definition includes tachycardias using a variety of reentry circuits. The reentry

circuits often occupy large areas of the atrium and are referred to as “macro‐

reentrant.” The classic type of atrial flutter (i.e. typical flutter) is dependent on the

cavotricuspid isthmus.

ACC/AHA/ESC 2003 Guidelines for the Management of Patients with Supraventricular

Arrhythmias (Blomstrom‐Lundqvist et al. 2003).

2

The typical form of atrial flutter is characterized by a sawtooth pattern of regular atrial

activation called flutter (ƒ) waves on the ECG, particularly visible in leads II, III, aVF, and

V1. If untreated, the atrial rate typically ranges from 240 to 320 beats per minute, with

ƒ waves inverted in leads II, III, and aVF and upright in lead V1. The direction of

activation in the right atrium may be reversed, resulting in upright ƒ waves in leads II,

III, and aVF and inversion in lead V1.

ACC/AHA/Physician Consortium 2008 Clinical Performance Measures for Adults with

Nonvalvular Atrial Fibrillation or Atrial Flutter (Estes et al. 2008).

1.1.1 History of Atrial Fibrillation

An irregular pulse, referred to as rebellious palpitations, delirium cordis and pulsus

irregularis perpetuus, was a cause of speculation by physicians since early times

(Silverman 1994). James Mackenzie (1907), a general practitioner in England, utilising

an ink‐writing polygraph to record and label jugular venous pulses, initially set about

deciphering normal and abnormal cardiac rhythms. His key observation that the

jugular "A wave" was lost in a patient who went from a normal to an irregular rhythm

provided the first insight into the mechanism of auricular fibrillation, as it was initially

called. The first electrocardiographic recordings of atrial fibrillation came shortly after

Einthoven invented the string galvanometer in 1901. He published a tracing from a

case of pulsus irregularis that showed QRS complexes with normal appearance; these

occurred irregularly but had too much background interference to permit the

identification of atrial activity (Einthoven 1906). Cushny and Edmunds (1907)

demonstrated a connection between irregularity of the pulse and atrial fibrillation.

3

Hering reporting on the electrocardiograms of patients in 1908, stated that one could

see no signs on the electrocardiogram of action of the auricles, and described “F

waves” in his recordings (Flegel 1995). These fine waves between ventricular beats

were considered to be responsible for the irregular rhythms whose pulse pressure

tracings had been published by MacKenzie. In 1909, Rothberger and Winterberg in

Vienna, and Lewis in England were the first to establish electrocardiographically that

auricular fibrillation was the cause of pulsus irregularis perpetuus (Silverman 1994).

Subsequently, Lewis (1910) noted from electrocardiographic studies that the R wave

was relatively normal in cases of irregular pulse and argued that ventricular contraction

must therefore originate from its usual starting point. He took the fine oscillations

between the R waves to be evidence of atrial activity throughout the cardiac cycle.

From detailed study of the chest leads, Lewis showed that these oscillations originated

from the atria rather than from the atrioventricular node as was thought by some

(Flegel 1995). Technical advances in ECG recording allowed signal amplification and

miniaturisation of the recording system (Ernestene and Levine 1928) which set the

stage for the subsequent path to understanding by way of electrophysiological

investigation of this frequently encountered clinical arrhythmia.

1.1.2 Epidemiology of Atrial Fibrillation

Atrial fibrillation is the most common chronic cardiac arrhythmia, estimated to affect

2.2 million people in the United States and 4.5 million people in the European Union

(Fuster et al. 2006). In Australia, it is a major health problem contributing to a

substantial proportion of the estimated 50,000 strokes a year (National Heart

4

Foundation of Australia 2004; Cadilhac et al. 2007). With ageing of the population and

improved survival after the occurrence of myocardial infarction and congestive heart

failure, Eugene Braunwald (1997) noted a doubling of admission rates for atrial

fibrillation between 1984 and 1994, and described atrial fibrillation as an emerging

epidemic. In Australia, the ICD‐10 principal diagnosis of “atrial fibrillation and flutter”

at discharge from public hospitals increased from 27,245 to 38,296 from 1999 to 2005

(Australian Institute of Health and Welfare 2008). Lifetime risks for development of

atrial fibrillation are 1 in 4 for men and women over the age of forty years and are still

high (1 in 6) even in the absence of antecedent congestive heart failure or myocardial

infarction (Lloyd‐Jones et al. 2004).

1.1.2.1 Incidence and Prevalence

The incidence and prevalence of atrial fibrillation varies between studies, most likely

related to study design and methods, but perhaps also to true regional variations.

Consistently across the studies, the incidence of atrial fibrillation increases with age

(Benjamin et al. 1994; Psaty et al. 1997; Miyasaka et al. 2006a). In the Cardiovascular

Health Study, the incidence of atrial fibrillation in the 65‐74 years age group was 18 per

1000 person‐years for men and 10 per 1000 person‐years for women. For the age

category 75‐84 years these numbers more than doubled to 43/1000 and 22/1000,

respectively (Psaty et al. 1997). The prevalence of atrial fibrillation is therefore highly

age‐related and although the age‐specific prevalence is higher in men (e.g. for ages 65‐

69 years: men 5.9%, women 2.8%), the gap narrows with women in the older age

categories. Furthermore, as women on average live longer than men, the absolute

number of women with atrial fibrillation actually exceeds that of men making this

5

disease of equal importance to both elderly men and women (Furberg et al. 1994). The

number of Australians with atrial fibrillation is estimated at 165,000, with numbers

expected to rise substantially due to ageing of the population together with the

increasing prevalence of atrial fibrillation with age (Eikelboom and Hankey 2004).

1.1.2.2 Associated Conditions

Key risk factors for the development of atrial fibrillation include increasing age,

hypertension, myocardial infarction, congestive heart failure and valvular heart disease

(Kannel et al. 1982a; Benjamin et al. 1994; Psaty et al. 1997). Data from the well known

Framingham cohort showed that, independent of other documented risk factors, each

decade of age increased the probability of developing atrial fibrillation by 2.1‐fold in

men and 2.2‐fold in women. In the same cohort, diabetes (men: 1.4‐fold; women 1.6‐

fold), underlying hypertension (men: 1.5‐fold; women: 1.4‐fold), past myocardial

infarction (1.4‐fold in men only), congestive heart failure (men: 4.5‐fold; women: 5.9‐

fold) and valvular disease usually affecting the mitral valve (men: 1.8‐fold; women: 3.4‐

fold) independently increased the likelihood of developing atrial fibrillation (Wolf et al.

1996). More recent data from the Framingham study reports that the development of

atrial fibrillation in patients with congestive heart failure is associated with an increase

in mortality of 2.7‐fold in men and 3.1‐fold in women (Wang et al. 2003). The Renfrew‐

Paisley study from the United Kingdom, found that middle‐aged men and women were

significantly more likely to develop atrial fibrillation during 25‐year follow‐up if they

had cardiomegaly (13‐fold increase), chronic bronchitis (2.2‐fold), left ventricular

hypertrophy (4.2‐fold), myocardial ischaemia (4.5‐fold), past history of stroke (3.9‐fold)

and glucose intolerance (3.1‐fold) (Stewart et al. 2001). Other reported causes include

6

pericarditis, alcohol, thyrotoxicosis, hypertrophic cardiomyopathy, myocarditis, severe

infection and disorders of the cardiac conduction system (Aberg 1968; Fuster et al.

2001; Stewart et al. 2001). Nevertheless, despite multiple risk factors for atrial

fibrillation, some of those with the condition are remarkable only for the absence of

overt cardiovascular disease or precipitating illness; so called “lone” atrial fibrillation

(Brand et al. 1985; Kopecky et al. 1987). Such patients younger that 65 years of age do

not have the increased risk of thromboembolism seen in atrial fibrillation overall

(Kopecky et al. 1987; Rostagno et al. 1995; Stewart et al. 2002). In fact, analysis of the

Framingham Study data of patients with lone atrial fibrillation revealed that after a

mean follow up of 25.2±9.5 years, cerebrovascular events had only occurred in

patients who had developed a least one risk factor for thromboembolism since

enrolment (Jahangir et al. 2007). Familial tendencies in lone atrial fibrillation have

been described and recently genetic linkage analysis studies have indicated

associations with specific chromosomal loci (Lai et al. 2003). Brugada et al. (1997)

reported the first monogenic cause for familial atrial fibrillation, implicating a gene on

chromosome 10. These studies, together with population studies of first degree

relatives (Ellinor et al. 2005), show there is growing evidence that genetic factors are

important in the pathogenesis of atrial fibrillation. Nevertheless, familial atrial

fibrillation has been shown to be genetically heterogeneous (Darbar et al. 2003;

Volders et al. 2007) and, unlike syndromes such as hypertrophic cardiomyopathy and

long QT syndrome, the role of genes in the vast majority of people who develop atrial

fibrillation probably comes second to the role of environmental factors and co‐morbid

disease.

7

1.1.2.3 Consequences of Atrial Fibrillation

Atrial fibrillation is a major cause of morbidity and mortality, increasing risk of death

(Benjamin et al. 1998; Vidaillet et al. 2002), congestive heart failure (Stewart et al.

2002), and embolic phenomena, including stroke (Wolf et al. 1991; Stewart et al.

2002). Approximately 15% of all strokes are attributable to atrial fibrillation (Wolf et al.

1987) and these are more debilitating than other types of strokes (Jorgensen et al.

1996). Independent of stroke, atrial fibrillation has been linked to cognitive

dysfunction although the mechanism by which this results is yet to be determined (Ott

et al. 1997; Miyasaka et al. 2007a). Congestive heart failure is not only a risk factor for

atrial fibrillation but can also be a consequence of atrial fibrillation. Therefore, whether

occurring in tandem or in succession these two conditions often co‐exist compounding

the adverse consequences of each in a “vicious cycle” (Cha et al. 2004). The incidence

of new congestive heart failure for a patient first diagnosed with atrial fibrillation has

been observed to be 44 per 1000 person‐years. Furthermore, the age‐ and sex‐

adjusted mortality risk for patients with atrial fibrillation following congestive heart

failure was estimated at 3.4 times that of those without heart failure (Miyasaka et al.

2006b). Atrial fibrillation confers an independent mortality risk, most recently shown in

a 21‐year community‐based study by the Framingham investigators. Patients with

newly diagnosed atrial fibrillation had a hazard ratio for dying of 2.08 over a mean of

5.3 years follow‐up, when compared with the general population of Minnesota.

Furthermore, this was markedly more dramatic (hazard ratio 9.62) in the first 4 months

from diagnosis, mainly from cardiovascular causes (Miyasaka et al. 2007b). In

Australians over the age of 60 years the relative mortality of patients with atrial

fibrillation (adjusted for confounding effects) is 1.92 for all causes and 3.78 for deaths

8

from stroke (Lake et al. 1989). Thus, atrial fibrillation is a disease that denotes

significant associated morbidity and mortality, with a large impact on both the

individual and the community as a whole. Importantly, these associated poor health

outcomes must be separated from the co‐morbid conditions frequently accompanying

atrial fibrillation. “Lone” atrial fibrillation has been defined as the absence of structural

heart disease or stroke based on history, physical examination, chest X‐ray, routine

blood chemistry, and trans‐thoracic as well as trans‐oesophageal echocardiography.

Additional exclusions to this group are coronary artery disease, significant pulmonary

disease, hypertension, hyperthyroidism and diabetes (Kopecky et al. 1987; Kumagai et

al. 1991; Jaïs et al. 2000). The survival for patients with “lone” atrial fibrillation has

been repeatedly shown to be comparable to the general population (Kopecky et al.

1987; Rostagno et al. 1995; Stewart et al. 2002). Although traditional teaching has

been that atrial fibrillation progresses from paroxysmal to persistent arrhythmia, a

recent Framingham study on the long‐term progression and outcomes of patients with

lone atrial fibrillation reported that of 71 patients with paroxysmal or persistent atrial

fibrillation, just 22 had progression to permanent atrial fibrillation over a mean follow‐

up of 25.2±9.5 years (Jahangir et al. 2007). Overall survival of the patients with lone

atrial fibrillation was similar to the survival of the age‐ and sex‐matched population.

Observed survival free of heart failure was slightly worse than expected but the risk for

cerebrovascular events was similar to that expected. Importantly, all patients who had

a cerebrovascular event had developed ≥1 risk factor for thromboembolism, thereby

were no longer strictly “lone” atrial fibrillation. While the association and

consequences of atrial fibrillation with all the factors described above may be clear,

the mechanisms by which these factors result in arrhythmia is not.

9

1.1.3 History of Atrial Flutter

The first published description of probable atrial flutter dates back to the nineteenth

century when McWilliam (1886) described regular, rapid excitations of the atrium in an

animal. Subsequently, Einthoven (1906) made an electrocardiographic recording of

atrial flutter. Distinction between atrial flutter and atrial fibrillation by characteristic

sawtooth waves in the inferior ECG leads followed (Jolly and Ritchie 1911). This

“common” type of atrial flutter was investigated by Lewis (1913) using a combination

of epicardial maps and ECG recordings from a canine model induced by rapid atrial

pacing. Continual activation of at least some part of the atrium resulted in flutter

waves seen in the surface ECG and the activation sequence was orderly, i.e. the wave‐

front circulated in either a cranio‐caudal or a caudo‐cranial direction in the right

atrium. From this pioneering experimental work it was concluded that atrial flutter was

due to intra‐atrial circus movement around the venae cavae (Lewis et al. 1921b).

Subsequent works that supported that atrial flutter was due to intra‐atrial reentry

included a crush injury model of this arrhythmia by creating a lesion between the

venae cavae (Rosenbleuth and Garcia‐Ramos 1947). Based on the epicardial maps, the

authors deduced that the reentry loop circled around the atrial crush lesion.

Interestingly, they also noted that when the crush lesion was extended from the

inferior vena cava to the atrioventricular groove, the arrhythmia disappeared and

could not be induced. This important finding suggested the true circuit may have

included the cavotricuspid isthmus. However, intra‐atrial macro‐reentry as the

mechanism of atrial flutter was not universally accepted. Goto et al. (1967) had shown

that aconitine caused abnormal automaticity at rapid rates in the rabbit atria. It was

thought that if the atrial site fired fast enough, either flutter (1:1 conduction) or

10

fibrillation (fibrillatory conduction because the atrial rate was too fast and 1:1

conduction could not be supported) occurred. Building on the work of Rosenbleuth

and Garcia‐Ramos, Frame et al. (1986) showed that the flutter reentry loop that

rotated around the tricuspid annulus could exist through creating a “Y” lesion in the

canine right atrium by extending the inter‐caval crush lesion to the right atrial free

wall. Subsequent experiments in various animal models and clinical studies have not

only confirmed that the mechanism of atrial flutter was due to intra‐atrial macro‐

reentry but also allowed the development of curative catheter ablation therapy. Of

particular importance are the studies which described techniques of manifest and

concealed entrainment, including identification of a site for catheter ablation (Waldo

et al. 1977; Inoue et al. 1981; Stevenson et al. 1995). Successful ablation of atrial

flutter by radiofrequency energy depended on the identification of a vulnerable,

critical zone in the reentrant circuit. Detailed analysis of intra‐operative mapping

studies of patients with persistent atrial flutter demonstrated that the narrowest part

of the circuit had relatively slow conduction and was localised to the low right atrium,

between the inferior vena cava and the tricuspid ring (Klein et al. 1986). Furthermore,

cryosurgical ablation of this critical region and its surrounding tissue prevented short‐

term recurrences of the arrhythmia. This was consistent with earlier work which noted

that when a crush lesion was extended from the inferior vena cava to the

atrioventricular ring, the arrhythmia could no longer be induced (Rosenbleuth and

Garcia‐Ramos 1947). Subsequent studies using direct current shocks to disrupt the

critical zone and eliminate the tachycardia supported the prospect that atrial flutter

could be permanently abolished by disruption of the isthmus (Saoudi et al. 1990).

However, one drawback of using direct current shock was that the shock itself could

11

convert atrial flutter. In the early 1990’s, two groups independently found that

disruption of the isthmus of the atrial flutter circuit could be carried out safely with

radiofrequency catheter ablation (Feld et al. 1992; Cosio et al. 1993). Subsequently,

using activation and entrainment mapping with guidance from intra‐cardiac

echocardiography, points in the right atrium that lay within or outside the circuit were

defined. The posterior barriers to conduction were identified as the crista terminalis

and the eustachian ridge (Olgin et al. 1995) and it was demonstrated that the tricuspid

annulus constitutes an anterior barrier constraining the reentrant wave front of human

anti‐clockwise atrial flutter (Kalman et al. 1996). Experiments in dogs show that the

atrial flutter circuit is critically dependent on a line of functional block extending to the

caval veins (Tomita et al. 2001). This has been confirmed in human studies showing the

crista terminalis as a line of functional block (Scaglione et al. 2000). Typical atrial flutter

therefore is a macro‐reentrant circuit in the right atrium with an anatomical barrier

anteriorly (the tricuspid valve) and a functional barrier posteriorly (the crista

terminalis).

1.1.4 Epidemiology of Atrial Flutter

Less data exists on the epidemiology of atrial flutter than atrial fibrillation, perhaps

because it is less common or because it is often included with atrial fibrillation in data

analysis, e.g. International Classification of Diseases, Tenth Revision. Certainly atrial

flutter shares many clinical risk factors with atrial fibrillation. Using the Marshfield

Epidemiological Study Area database of 58,820 residents, Granada et al. (2000)

ascertained all new cases of atrial flutter over a 4 year period from 1991. One‐hundred

12

and eighty‐one new cases of atrial flutter were diagnosed for an overall incidence of

88/100,000 person‐years. Incidence rates ranged from 0.005% in those 50 years old to

0.6% in subjects older than 80. Atrial flutter was 2.5 times more common in men, 3.5

times more common in subjects with heart failure and 1.9 times more likely for

subjects with chronic obstructive pulmonary disease. Among those with atrial flutter,

16% were attributable to heart failure and 12% to chronic obstructive lung disease.

Three subjects (1.7%) without identifiable predisposing risks were labelled as having

“lone atrial flutter”. When these figures are applied to the entire population of the

United States, an estimated 200,000 new cases of atrial flutter occur there annually.

With regard to mortality, atrial flutter appears more benign than atrial fibrillation, at

least initially. Vidaillet et al. (2002) found that for 6 months after diagnosis, mortality

was 2.5 times higher among those with atrial flutter (5%) than among controls (2%),

but 3.4 times lower than among patients who had atrial fibrillation with or without

atrial flutter (17%). After adjusting for baseline characteristics however, early mortality

was about seven times higher for patients with atrial fibrillation than for controls. With

longer follow‐up (mean 3.6 years), overall mortality was similar for patients in all

arrhythmia subgroups – 41% for atrial flutter, 45% for atrial fibrillation and 47% in

patients with both atrial flutter and atrial fibrillation – as compared with 22% in

controls. All three groups of patients with atrial fibrillation or atrial flutter had an

approximately two‐fold increased risk of mortality compared with controls in models

that adjusted for other risk factors. Mortality among patients with atrial fibrillation

trended somewhat higher (1.4‐fold) than among patients with atrial flutter alone.

13

1.2 The Electrophysiological Basis of Atrial Fibrillation

1.2.1 The Multiple Heterotopous Centres Theory

Initial theories of the mechanism of atrial fibrillation proposed that each heart fibre

becomes independently rhythmic and that each is a focus of its own impulse formation

as a result of increased excitability; the multiple heterotopous centres theory. Lewis

and Schleiter (1912) reasoned that activity from one or more heterogeneous centres

would account for single premature beats, for regular tachycardias and, ultimately, for

the completely incoordinate activity seen in auricular fibrillation. However, two unique

observations were made by Garrey (1914). Firstly, when a fibrillating chamber was cut

into four pieces, each fragment continued to fibrillate; this could not have happened if

the fragments were dependent on a single focus. Secondly, although portions of

isolated heart muscle were capable of fibrillating, a critical amount of muscle mass was

necessary. These observations made the multiple heterotopous centres theory

untenable and a competing theory of “circus movement” was proposed in which a

circuit of muscle could propagate a wave‐front round and round indefinitely, provided

the length of muscle exceeded the wave‐front length (Lewis et al. 1921b).

1.2.2 The Multiple Wavelet Hypothesis

The circus movement theory was developed further to explain the complexity of

experimental findings; these developments included the concept of a single "mother"

ring that propagated the arrhythmia and gave rise to "daughter" rings and, finally, to

multiple independent rings (Moe 1962). This theory of multiple circus movement, or

wavelets, dominated the thinking on the mechanism of atrial fibrillation for many

14

years. Moe hypothesised that a grossly irregular wave‐front becomes fractionated as it

divides around islets or strands of refractory tissue and each of the daughter wavelets

could then be considered independent offspring; the multiple wavelet hypothesis.

However, prior to this, Prinzmetal et al. (1950) had showed that premature systoles,

paroxysmal tachycardia, atrial flutter and atrial fibrillation of the left auricle

(investigated in over 200 dogs by high speed cinematography, cathode‐ray

oscillography and multiple‐channel electrocardiography) were of unitary origin and

suggested that all may occur from one ectopic focus. They categorically stated there is

no circus movement and claimed their conception of the auricular arrhythmias

simplified the understanding of the mechanism. Subsequent to Moe’s theory however,

Allessie et al. (1977) demonstrated reentrant circuits in isolated rabbit atria by using

precisely timed premature impulses to produce sustained tachycardia. Of particular

interest as an explanation of atrial fibrillation is the "leading circle” theory in which the

initiation of reentry takes place because of non‐uniform refractory periods in atrial

fibres in close proximity to one another. The initiating impulse conducts in fibres with

short refractory periods and is blocked in those with longer ones, forming the

conditions for reentry before the impulse has died out. Impulses circulate around a

central area that is kept refractory, and thereby blocked, by centripetal wavelets

arriving from all sides. The size of the circle is determined by the recovery time of the

tissue forming the circuit, because tissue on both sides is kept depolarised by the

leading circle. Further evidence to support this hypothesis came about when Cox et al.

(1991) demonstrated in an experimental dog model, with supplementary evidence

from intra‐operative clinical observations in humans, that macro‐reentrant circuits

within the atrial myocardium were responsible for the entire spectrum of atrial

15

arrhythmias. Both the experimental study and the clinical study demonstrated that

multiple wave fronts, non‐uniform conduction, bi‐directional block, and macro‐

reentrant circuits occur during atrial fibrillation. Subsequent work by Allessie’s group

on conscious dogs underscored the importance of slowed conduction and shortened

refractoriness — a short wavelength, the distance travelled by the depolarisation wave

during the refractory period — in determining the onset of atrial fibrillation during

rapid electrical stimulation (Rensma et al. 1988). Moe’s original theory of multiple

wavelets was lent considerable weight by the clinical observation that chronic atrial

fibrillation could be cured in some patients by the placement of multiple surgical

lesions (the maze procedure) to compartmentalise the atria into regions presumably

too small to sustain such wavelets (Cox et al. 1993). Intra‐operative mapping of human

atrial fibrillation found pivot points where the wavelets turn around at the end of lines

of functional block, illustrating the complexity of atrial activation (Konings et al. 1994).

1.2.3 Localised Sources in the Initiation and Maintenance of Atrial Fibrillation

At the time that surgical treatments of atrial fibrillation were being advanced, the first

direct evidence that the pattern of circulating wavelets during atrial fibrillation in

humans is not entirely random was published (Gerstenfeld et al. 1992). The hypothesis

that local atrial activation should be influenced by the constant anatomy and the

receding tail of refractoriness from the previous activation was proven by

demonstrating that a tendency for wave fronts to follow paths of previous excitation

(often termed "linking") was transiently present for the majority of patients with atrial

fibrillation. Further evidence that atrial fibrillation is not random comes from more

16

recent studies showing repetitive periodic activity demonstrating temporal and spatial

stability in the left atrium, particularly the pulmonary veins and posterior regions

(Skanes et al. 1998; Mansour et al. 2001). Analysis of the activation frequency at these

sites is highly suggestive of localised reentry or “rotors” which have been proposed as

drivers of the fibrillatory process (Jalife et al. 1998; 2002). Mansour (2001) reported

that wave‐fronts propagated from left to right in 80% of sheep atria and that right‐to‐

left propagation occurred in a significantly smaller percentage of cases. The cycle

length of left atrial sources determined the dominant peak in the frequency spectra in

atrial fibrillation and ablation of inter‐atrial connections reduces right atrial signal

frequency but not left. Human studies have also suggested localised reentry as a

driving mechanism of atrial fibrillation and are discussed further in Section 1.3.2. After

the leading circle wavelet theory had held sway for decades, Scherf (1947) revived the

focus theory by applying aconitine focally to the atrium, and in this way could induce

atrial fibrillation. More recently, research has similarly focussed on identifying local

triggers for atrial fibrillation. It has been observed that atrial fibrillation can ensue from

the degeneration of other atrial tachycardias such as atrial flutter or accessory

pathway mediated rhythms, and in fact ablation of such pathways can reduce the

incidence of atrial fibrillation (Haïssaguerre et al. 1992). Focal atrial tachycardia is

thought to arise from abnormal automaticity of cardiac cells and shows a predilection

for sites in close proximity to anatomical structures such as the crista terminalis (Chen

et al. 1994; Kalman et al. 1998), the atrial septum (Chen et al. 2000; Marrouche et al.

2002), the coronary sinus (Volkmer et al. 2002; Badhwar et al. 2005; Kistler et al.

2005), the atrioventricular annuli (Morton et al. 2001; Matsuoka et al. 2002; Kistler et

al. 2003b; Hachiya et al. 2005) and the pulmonary veins (Kistler et al. 2003a; Hachiya et

17

al. 2005). Cells around the tricuspid annulus with nodal characteristics have been

found (McGuire et al. 1996). Although histologically similar to atrial cells they

resembled nodal cells in their cellular electrophysiology, response to adenosine, and

connexin expression which suggests that nodal‐like cells capable of sustaining ectopic

activity exist in these structures where focal tachycardia has been found to originate.

Perhaps the most important source of ectopy for the initiation of atrial fibrillation is

that arising from the pulmonary veins (Haïssaguerre et al. 2000a). In a landmark study

94% of the 69 ectopic foci found to initiate atrial fibrillation in 45 patients with drug‐

refractory episodes were within the pulmonary veins (Haïssaguerre et al. 1998).

Furthermore, in the same study, radiofrequency ablation of these same foci kept 62%

of these subjects free of atrial fibrillation at a mean of eight months follow up. A

further study targeting the pulmonary veins by a slightly different technique of wide

circumferential ablation with good long term success lends further weight to the role

of the pulmonary veins in the genesis and maintenance of atrial fibrillation (Pappone

et al. 2000). Anatomical investigation of the pulmonary veins reveals sleeves of atrial

myocardium extending into the proximal veins a variable distance (Nathan and Eliakim

1966; Moubarak et al. 2000), the length of which is greatest in the superior veins (Saito

et al. 2000). Myocardial bundles can be seen to cover the complete pulmonary vein

circumference in 54% of veins examined 1cm from the ostium (Weiss et al. 2002).

Fibres of muscle in these ostial regions are variably arranged with one or more layers

of longitudinal, circular, oblique or spiral fibre orientation interwoven with non‐

conducting connective tissue, perhaps explaining the observation of the non‐uniform

breakthrough pattern of pulmonary vein activity into left atrial tissue (Haïssaguerre et

al. 2000b). Zones of activation delay have been observed in canine pulmonary veins

18

and correlated with abrupt changes in fascicle orientation, perhaps facilitating reentry

and arrhythmias (Hocini et al. 2002). Reports of pulmonary vein aneurysm being

implicated as a trigger for atrial fibrillation (Yamane et al. 2000) lend further weight to

the importance of vein structure in the electrical output to the left atrium. There is

also evidence to support the existence of important differences in membrane channel

distribution between the pulmonary veins and the left atrium (Melnyk et al. 2005) as

well as greater variation in myocyte size and fibrosis in the pulmonary veins of patients

with atrial fibrillation (Tagawa et al. 2001). The stage may then be set for reentry to

occur within or around the pulmonary veins when triggered by short‐coupled

“automatic” discharges (Shah et al. 2001). Study of human embryos with

immunohistochemical staining of the conduction system has shown specialised

conduction tissue at areas known to demonstrate abnormal automaticity in the adult

(Blom et al. 1999). Electron microscopy of autopsy specimens has suggested the

presence of P cells, transitional cells and Purkinje cells in the pulmonary veins (Perez‐

Lugones et al. 2003). Very recently, histological investigation of pulmonary vein sleeves

has identified “interstitial cells of Cajal”; cells known to act as pacemakers in the gut

which undoubtedly have the potential to act as a pacemaker through their expression

of ion channels with If properties (Morel et al. 2008). Clearly, these “triggers” found in

the pulmonary veins and other sites are important in the initiation of atrial fibrillation,

as backed up by evidence to show re‐initiation of atrial fibrillation following

cardioversion may be a result of local discharge from these areas (Lau et al. 1999;

Haïssaguerre et al. 2000c). Why these areas should be predisposed to initiating atrial

fibrillation and how their independent discharge should bring about wavelet reentry

has been the subject of intense investigation.

19

1.2.4 The Venous Wave Hypothesis

Building on work gone before, Haïssaguerre et al. (2004b) proposed the venous wave

hypothesis. This was an alternative hypothesis to the classical multiple wavelet

mechanism proposed by Moe that in at least some patients with paroxysmal atrial

fibrillation the pulmonary vein is the source of fibrillatory activity that maintains the

atria in fibrillation. Studies have shown that the pulmonary veins are capable of

automaticity and triggered activity under conditions of rapid atrial pacing or congestive

heart failure (Chen et al. 2001; Okuyama et al. 2003). The pulmonary veins of patients

with atrial fibrillation have been shown to have distinctive electrophysiological

properties when compared to those of patients without arrhythmia, and also to the

atria to which they are attached (Jaïs et al. 2002). The specific architecture of the

muscular sleeves within the pulmonary veins of dogs may facilitate reentry (Hocini et

al. 2002), and indeed reentry has been observed within human pulmonary veins at

multi‐electrode basket studies (Kumagai et al. 2004). Pulmonary veins identified as

initiators of arrhythmia demonstrate periods of short cycle length during ongoing atrial

fibrillation (O'Donnell et al. 2002). Isolation of such pulmonary veins renders the atria

less inducible to atrial fibrillation and therefore suggests a dynamic interplay occurring

between the atria and pulmonary veins (Oral et al. 2002). Chen et al. (1999) reported

longer refractory periods at the proximal portions of the pulmonary veins and this

distal‐to‐proximal gradient has been confirmed by others (Kumagai et al. 2004). This

may be somewhat protective against more distal rapid activation being conducted to

the atria, however this difference was attenuated with isoprenaline emphasising the

dynamic role of the autonomic system in the electrophysiology of the pulmonary veins

(Chen et al. 1999). Rapid pacing of isolated pulmonary veins following ablation can

20

induce local sustained reentry and features of decremental conduction within the

paced vein (Takahashi et al. 2003). Arora et al. (2003), performing optical mapping

studies of normal canine pulmonary veins, demonstrated that these structures possess

both anisotropic conduction and repolarisation heterogeneity. With extrastimulus

testing they observed regions of uni‐directional conduction block and slowed

conduction that initiated leading circle reentry which became sustained with

isoprenaline. Pulmonary vein isolation has been shown to produce a progressive

prolongation of the atrial fibrillation cycle length, varying in extent from vein to vein

and between individuals, culminating in arrhythmia termination in most patients after

a significant cumulative increase in the cycle length (Haïssaguerre et al. 2004a). The

distinctive electrophysiological properties of the pulmonary veins of patients with

atrial fibrillation have been shown to be emphasised by amiodarone therapy and may

be responsible for heterogeneous alteration of pulmonary vein electrophysiology

(Rostock et al. 2005). Recently, Rostock et al. (2008) have demonstrated differential

rate‐adaptation of the electrophysiological properties of the pulmonary veins in

patients without a history of atrial fibrillation. They observed that the effective

refractory period (ERP) of the pulmonary veins was higher than the ERP in the atria at

baseline, and that the shortening of ERP in response to 15 minutes of atrial fibrillation

was greater in the pulmonary veins than in the atria. This suggests that the changes

within the pulmonary veins promote the maintenance of atrial fibrillation to a greater

degree than the atria. Thus, there potentially exists within the pulmonary veins alone

the electrophysiological milieu necessary for both the initiation and maintenance of

atrial fibrillation. This hypothesis is further supported by evidence of continued

fibrillatory activity seen in disconnected pulmonary veins (Willems et al. 2003) and the

21

common observation of slow dissociated pulmonary vein activity. The venous wave

hypothesis therefore suggests that pulmonary veins have heterogeneous

electrophysiological properties capable of sustaining reentry and that these structures

are implicated not only as a trigger of atrial fibrillation in patients with the appropriate

substrate, but also as a source of “venous waves/drivers” that are capable of

maintaining the atria in fibrillation.

1.2.5 Rotors and Activation Gradients

Localised regions of high‐frequency activity demonstrating spatio‐temporal periodicity

have been suggested to act as drivers of atrial fibrillation (Skanes et al. 1998; Jalife et

al. 2002; Everett et al. 2006; Lin et al. 2006). Jalife et al. (1998) proposed the theory of

“mother rotors” as drivers of atrial fibrillation, with fibrillatory activity of nearby

myocardium unable to maintain 1:1 conduction with the rapid site. Evidence for

organisation of atrial fibrillation has come from animal optical mapping studies which

have used spectral analysis to demonstrate repetitive high frequency activity during

atrial fibrillation, consistent with the hypothesis of “driving” rotors (Kalifa et al. 2006).

Similar findings have also been reported in humans (Wu et al. 2002; Lazar et al. 2004;

Sahadevan et al. 2004; Sanders et al. 2005a; Haïssaguerre et al. 2006b). Gradients of

activation from the pulmonary veins to the remaining atrium of patients with

paroxysmal atrial fibrillation have been reported (Sanders et al. 2006). However, such

specific patterns of activation gradient are less conclusive in persistent atrial

fibrillation. Retrospective analysis of ablation of paroxysmal atrial fibrillation shows

that atrial fibrillation cycle length prolongation and eventual termination occurs at

22

sites of high frequency activation (Sanders et al. 2005a); this has not been shown for

persistent atrial fibrillation. Further discussion of the identification of high frequency

activation sites and their role in the atrial substrate is discussed in Section 1.3.5.

1.2.6 Summary

Since the first descriptions of atrial fibrillation, our understanding of the nature of this

rhythm has expanded and the explanations for its occurrence evolved. From the initial

Multiple Heterotopous Centres Theory, to the Multiple Wavelet Hypothesis, the more

recent Venous Wave Hypothesis, and the proposed concept of Rotors driving atrial

fibrillation, it is apparent that complex and varied mechanisms are at play in the

initiation and maintenance of this rhythm. It is probable that both focal activity and

reentry are responsible for the maintenance of atrial fibrillation. However, these

mechanisms may not have to be present in the same person at the same time (Wu et

al. 2004). Changes in electrophysiological characteristics of the myocardium could

convert multiple wavelet reentry atrial fibrillation into focal discharge atrial fibrillation

and vice‐versa. Alternatively, a partnership between focal discharge and reentrant

mechanisms might be necessary for maintenance of atrial fibrillation. Much is still to

be learnt about the electrophysiological basis of atrial fibrillation.

1.3 Triggers and Substrate in Atrial Arrhythmia

The concept of triggers, initiators and perpetuators advanced by Allessie et al. (2001)

requires that for atrial fibrillation to exist it must be initiated by some trigger and then

maintained by favourable electroanatomical properties, or substrate.

23

1.3.1 Focal Triggers

Abnormal impulse initiation may be divided into automaticity and triggered activity.

Automaticity is the property of cardiac cells to undergo spontaneous diastolic

depolarisation and initiate an electrical impulse in the absence of external electrical

stimulation (Peters et al. 2000). This is a known property of a variety of cardiac tissues

including the sinus node and subsidiary pacemaker sites and is thought to be due to

gradual inward net current (If) (Brown et al. 1979). Triggered activity is a term used to

describe impulse initiation dependent on after‐depolarisations. After‐depolarisations

are oscillations in membrane potential that follow the upstroke of an action potential

(Cranefield 1977). Delayed after‐depolarisations occur following the repolarisation of

an action potential. This can cause triggered activity if the after‐depolarisation reaches

threshold for a subsequent action potential. The most widely recognised cause of

delayed after‐depolarisation triggered activity is digitalis (Ferrier and Moe 1973). In

contrast, early after‐depolarisations occur prior to the repolarisation of an action

potential. These tend to delay the repolarisation phase and are important in the

mechanism of the congenital long QT syndromes. Mutated genes such as SCN5A,

KVLQT1 and HERG prolong action potential duration and allow early after‐

depolarisations (Roden et al. 1996). Evidence supporting focal sources in promoting

atrial fibrillation comes from studies characterising atrial tachycardia late after ablation

for atrial fibrillation. Haïssaguerre et al. (2005b) reported on 55 atrial tachycardias of

which 20 were characterised as focal in origin, 27 as macro‐reentrant and the

remaining 8 (localised to the atrial septum) displayed mixed mechanistic

characteristics. Gerstenfeld et al. (2005) reported a small series of 5 patients in whom

a focal atrial tachycardia was present following segmental pulmonary vein isolation.

24

Arentz et al. (2007) performed a sophisticated basket‐mapping study of the pulmonary

veins to characterise spontaneous activity initiating atrial fibrillation and found that

the first ectopic activity of all events was of focal origin. Nevertheless, some

tachycardias from the distal pulmonary veins of short cycle length demonstrated

continuous activity throughout a significant proportion of the tachycardia cycle length

favouring some reentrant mechanism occurring subsequently. Despite the highly

sophisticated mapping techniques employed there remains some methodological

problems in assessing the mechanism of tachycardia in pulmonary veins; entrainment

mapping is difficult at high rates, pacing manoeuvres may terminate tachycardia,

capture of the pulmonary vein without capture of adjacent structures is an issue, and

rapid pacing of itself induces fractionated electrograms (Willems and Rostock 2007).

Markowitz et al. (2007) have demonstrated that adenosine can assist in the

differentiation between focal and reentrant tachycardia, and furthermore micro‐

reentrant circuits can also be distinguished from automatic focal atrial tachycardia by

this means. This also suggests that some focal sources previously thought to be due to

automaticity or triggered activity may in fact be due to localised reentry.

1.3.2 Localised Reentry

The requirements for reentry are:

(i) uni‐directional conduction block,

(ii) a core of inexcitable tissue around which the wave‐front propagates, and

(iii) the maintenance of excitable tissue ahead of the propagating wave‐front

(the “excitable gap”) (Peters et al. 2000).

25

The excitable gap is facilitated by slowing of conduction and/or shortening of the

refractory period. The wavelength (in metres) is the product of the conduction velocity

(in metres/second) and the refractory period (in seconds). This is critical to the

development of atrial reentry; a shorter wavelength requires less atrial mass to sustain

reentry. Or to put it another way, a given atrial mass can sustain more reentrant

circuits if the wavelengths are shorter. It is important to realise that the above

requirements may be anatomical or functional. The core of inexcitable tissue may be

non‐myocardial tissue (such as a valve) or an area of functional block (such as the

crista terminalis) (Yamashita et al. 1992; Tai et al. 1998). The excitable gap will depend

on the refractory period of the circuit, itself highly influenced by drugs, hormonal and

autonomic factors. Sanders et al. (2005b) demonstrated local reentry by performing

endocardial mapping with a multi‐spline catheter during focal atrial tachycardia.

Features such as entrainment and a large proportion of cycle length observable within

the mapped area were able to be established. Haïssaguerre et al. (2005b) have

reported similar observations during atrial fibrillation organised by prior ablation.

Indeed, localised reentry can be mapped to sites where the greatest impact of ablation

on the atrial fibrillation cycle length has been shown to occur; the pulmonary veins,

the fossa ovalis, the coronary sinus and the base of the left atrial appendage. Although

these may be iatrogenic from the prior ablation procedure, the fact that they share

common sites with critical regions maintaining atrial fibrillation is evidence that these

areas of localised reentry are integral to the initiation and maintenance of arrhythmia.

This group subsequently were able to demonstrate features of small circuits consistent

with local reentry by simultaneous mapping with a multi‐spline catheter during the

26

latter stages of the initial ablation procedure for atrial fibrillation when periods of

relative organisation of electrograms can be seen (Haïssaguerre et al. 2006a).

1.3.3 Anatomical Structures with Site‐specific Conduction Properties

Integral to the theories of the electrophysiological basis of atrial fibrillation is the

concept of heterogeneous conduction properties. There are regions within the atria

where predictable alterations in these properties are seen.

1.3.3.1 Right Atrium

The crista terminalis has unique conduction properties; specifically the gap junctions

are located at the poles of muscle cells resulting in a higher degree of “end‐to‐end”

connections than other myocardial tissue (Saffitz et al. 1997). This confers a high

degree of anisotropy to this structure, resulting in conduction velocities 10 times

greater in the longitudinal than the transverse direction (Saffitz et al. 1994). This

feature of preferential conduction has been studied for some time (Spach et al. 1981)

and had led some investigators to conclude that specialised conduction exists within

the atria, despite no histological evidence of Purkinje‐equivalent fibres (Liebman

1985). Numerous studies have explored the functional conduction properties of the

crista terminalis (Olgin et al. 1995; 1996; Tai et al. 1998; 2004). In addition, site‐specific

conduction delays in the low right atrial myocardium and the presence of non‐uniform

anisotropic characteristics of the posterior triangle of Koch have been suggested as

critical substrate for induction of atrial fibrillation (Papageorgiou et al. 1996). This area

of the atrium has been shown to have complex electrophysiological properties and

27

multiple inputs to the atrioventricular node (McGuire et al. 1993) which may account

for its unique conduction characteristics.

1.3.3.2 Inter‐atrial Connections

The inter‐atrial connections are regions where the electrical wave‐front traverses from

right to left during sinus rhythm. The most important and well studied of these is

Bachmann’s bundle which is located superiorly on the anterior surface of the inter‐

atrial septum. This connection has been demonstrated in animal experiments to have

differing conduction properties than other atrial myocardium (velocity of conduction is

faster and ERP is higher), representing a potential substrate for reentry to occur

(Platonov 2007). In humans, left atrial mapping studies have demonstrated

Bachmann’s bundle as well as posterior and inferior connections in the vicinity of the

coronary sinus (De Ponti et al. 2002; Markides et al. 2003; Betts et al. 2004; Lemery et

al. 2007). Most of these studies have been conducted in patients with atrial fibrillation

during sinus rhythm and demonstrate Bachmann’s bundle as critical, however for atrial

ectopic beats occurring from the posterior left atrium it is easy to propose how the

inferior connections may contribute to initiation of arrhythmia. In addition, fibro‐fatty

degeneration of Bachmann’s bundle has been reported as common among patients

with a history of atrial fibrillation (Becker 2004). Recently, a study of 84 post‐mortem

hearts has confirmed inter‐atrial muscular connections present anteriorly at

Bachmann’s bundle, posteriorly between right pulmonary veins and inferiorly between

the coronary sinus and right inferior pulmonary vein (Platonov et al. 2008). Several

investigators have used the knowledge of the anatomy of inter‐atrial connections to

guide placement of an atrial pacemaker lead with the aim of shortening atrial

28

activation and thereby reducing atrial fibrillation (Bailin et al. 2001; Padeletti et al.

2001; Hermida et al. 2004; de Voogt et al. 2007). The results of these studies are

mixed, perhaps due to varying positions of the pacing lead and accompanying anti‐

arrhythmic pacing algorithms, as well as heterogeneity within the atrial fibrillation

population itself. Nevertheless, it remains a promising area of research for patients

with sick sinus syndrome, a group in whom atrial fibrillation is common.

1.3.3.3 Left Atrium

In the left atrium, several structures (besides the pulmonary veins; see Sections 1.2.3

and 1.2.4) have been investigated for their site‐specific properties. The left lateral

ridge (the infolding of myocardium between the left pulmonary vein antra and the

appendage) is an area of interest due to the close proximity of the vein of Marshall and

the ganglionic plexus epicardially (Cabrera et al. 2008). It is also a challenging area for

ablation, due to its narrow superior aspect making catheter stability difficult. The

septopulmonary bundle is of recent interest as a left atrial “crista‐equivalent”.

Markides et al. (2003) demonstrated by non‐contact mapping a vertical line of

functional block in the posterior left atrium corresponding to a region of abrupt change

in fibre orientation seen in 10 post‐mortem hearts. High‐resolution mapping in a sheep

model has demonstrated wave‐break occurring at the regions along the

septopulmonary bundle that showed sharp changes in fibre direction and abrupt

increases in myocardial thickness (Klos et al. 2008). Waves propagating from the

pulmonary veins into the posterior left atrium triggered reentry and atrial fibrillation

by breaking at boundaries along this bundle. Roberts‐Thomson et al. (2008) have

described a vertical line of functional conduction delay in a consistent anatomical

29

location in the posterior left atrium of patients with structural heart disease, most

markedly with atrial dilatation. This line facilitates circuitous wave‐front propagation,

suggesting a potential role in arrhythmogenesis.

1.3.4 Complex Fractionated Atrial Electrograms

One of the characteristic features of the underlying substrate of atrial fibrillation is the

presence of complex fractionated atrial electrograms (CFAE). CFAE observed during

intra‐operative mapping of human atrial fibrillation are found mostly in areas of slow

conduction and/or at pivot points where the wavelets turn around at the end of arcs of

functional block (Konings et al. 1994). Results from this pivotal publication by the

laboratory of Allessie and colleagues were confirmed by later studies in which the

morphology of electrograms during atrial fibrillation reflected the occurrence of

various specific patterns of conduction (Konings et al. 1997). The authors postulated

that electrogram morphology might be used to identify regions with structural

conduction disturbances involved in the perpetuation of atrial fibrillation. Prior to this,

Cosio et al. (1986) had identified complex electrograms as important in the

determination of abnormal conduction in atrial flutter. Jaïs et al. (1996) initially

described regional bi‐atrial disparities in endocardial fractionation in paroxysmal atrial

fibrillation and found that complex electrical activity was more frequently found in the

septal and posterior regions of both atria, with more regular activity seen in the

trabeculated atria. These findings led Nademanee et al. (2004) to hypothesise that if

CFAE were to be selectively eliminated by catheter ablation, wavelet reentry should

stop, thereby preventing atrial fibrillation perpetuation. The earlier observations that

30

led to the multiple wavelet theory as the underlying mechanism for atrial fibrillation

(random reentry with meandering wavelets) is at odds with the observation of site‐

specific reentry. Possible explanations include the linking phenomenon (Gerstenfeld et

al. 1992), which may explain maintenance of the direction of wavelet propagation

during atrial fibrillation, and the existence of adverse electrical remodelling with

repeated episodes of atrial fibrillation (Wijffels et al. 1995). By targeting sites of CFAE

for ablation, Nademanee et al. (2004) reported 70% of a balanced paroxysmal and

persistent cohort were arrhythmia‐free off medication one year following a single

ablation procedure. Furthermore, with a mean of 1.24 ablation procedures and a small

minority still on anti‐arrhythmic drugs, 91% were free of atrial fibrillation. The inter‐

atrial septum was the most common site for CFAE in this study; other common sites

were the pulmonary veins, left atrial roof and proximal coronary sinus. It is important

to note that while many previous studies have used an epicardial approach where

areas such as the septum, pulmonary veins and coronary sinus os are not accessible

(Cox et al. 1991; Gerstenfeld et al. 1992), these were precisely the sites found to

contain CFAE at endocardial mapping. CFAE may be seen more frequently at sites of

known site‐specific conduction abnormalities (Roberts‐Thomson et al. 2008). In

addition, an association between CFAE and atrial fibrillation cycle length was

demonstrated by Rostock et al. (2006) who observed a shortening in cycle length

preceding the development of maximum fractionation at 91% of mapped sites,

suggesting a functional nature to these electrograms. More recently, Nademanee et al.

(2008) have published long‐term outcomes for high‐risk patients ablated by this

technique. For a mean of 1.68 procedures per patient, 81% of these challenging cases

were in sinus rhythm at an average of 2.3 years follow up. Just 13% of those free of

31

atrial fibrillation required anti‐arrhythmic agents to maintain sinus rhythm. However,

these results have not been reproduced by operators seeking to emulate this success.

Although most investigators have incorporated CFAE‐targeted ablation into a multi‐

faceted ablation approach (Takahashi et al. 2007; Porter et al. 2008; Verma et al.

2008), Oral et al. performed solely CFAE‐targeted ablation in 100 patients with chronic

atrial fibrillation and reported modest success of 33% maintenance of sinus rhythm

without drugs at 14 months mean follow‐up. Perhaps one reason for the lack of

reproducible outcomes by this method is that operators have been reliant on

subjective analysis of CFAE classification as guided by the published definitions.

Definitions for CFAE are plentiful and varied, and include:

(i) electrograms with an FF interval <100ms or continuous activity (Jaïs et al.

1996),

(ii) fractionated electrograms composed of ≥2 deflections, perturbation of the

baseline with continuous deflection of a prolonged activation complex, or atrial

electrograms with a cycle length ≤120ms (Nademanee et al. 2004),

(iii) fractionated potentials with ≥3 deflections from the isoelectric line or

continuous activity as their criteria for CFAE (Rostock et al. 2006), and

(iv) electrograms with a cycle length ≤120ms or shorter than the coronary sinus, or

those that were fractionated or displayed continuous electrical activity (Oral et

al. 2007).

It is perhaps the variability of these descriptions that accounts for some of the

inconsistent results reported across different centres using CFAE‐targeted ablation.

Takahashi et al. (2008) attempted to define the characteristics of electrograms which

32

identified successful ablation by measuring the effect of ablation on change in atrial

fibrillation cycle length. They found that electrograms with a temporal activation

gradient of >70ms or continuous activity indicated favourable sites for effective

ablation. Algorithms for quantification of electrogram fractionation have become

commercially available and the use of such algorithms to complement established

ablation strategies should aim to target critical sites maintaining fibrillation, while

sparing bystander atrium. The first publication using an automated fractionation

detection tool to describe ablation outcomes has taken advantage of the reproducible

nature of automated algorithms allowing it to guide ablation in a consistent fashion

across three centres to demonstrate an incremental success rate over historical

controls (Verma et al. 2008). While there is mounting evidence to suggest that

mapping CFAE by automatic techniques identifies critical ablation sites, the

effectiveness of CFAE‐targeted ablation is yet to be widely demonstrated.

Nevertheless, a consistent approach to characterise fractionation is the first step

towards establishing a reproducible method on which trials designed to address the

efficacy of CFAE‐guided ablation may rely.

1.3.5 Dominant Frequency

Analysis of electrograms within the time domain has led to identification of features

such as automatic activity and complex fractionation, however due to the apparently

chaotic nature of electrical activity during atrial fibrillation it is difficult to see

underlying patterns of depolarisation that may be important in the substrate

maintaining atrial fibrillation. This has led to the analysis of electrograms in the

33

frequency domain by fast fourier transform. Using this technique, information about

local activation may be distilled from an electrogram otherwise too complex or noisy

to be appreciated in the time domain. The atrial electrograms obtained during

fibrillation are often such electrograms and hence fast fourier transform “unlocks” that

information for analysis. The fast fourier transform attempts to reduce a complex

electrogram to its constituent sine wave components and plots the frequency of these

waves against the amplitude of the wave power to represent the electrogram in the

frequency domain. The frequency value demonstrating peak power in this spectrum is

termed the dominant frequency (DF). A strong correlation has been shown

experimentally between DF and atrial rotor frequency observed by optical mapping

(Sarmast et al. 2003). Sites demonstrating high DF have been proposed as “drivers” of

atrial fibrillation (Jalife et al. 1998; Berenfeld et al. 2000; Sanders et al. 2005a).

Evidence for organisation of atrial fibrillation from optical mapping in animal studies

has used spectral analysis to demonstrate repetitive high frequency activity with

temporal and spatial periodicity during atrial fibrillation (Skanes et al. 1998; Kalifa et al.

2006). Similar findings by alternative, lower‐resolution techniques have been reported

in humans (Wu et al. 2002; Lazar et al. 2004; Sahadevan et al. 2004; Sanders et al.

2005a). Higher DF has been reported in the left atrium of patients with paroxysmal

atrial fibrillation (Lazar et al. 2004; Lin et al. 2006). Furthermore, the maximal DF has

been reported in the pulmonary veins in patients with paroxysmal but not persistent

atrial fibrillation (Sanders et al. 2006). Persistent atrial fibrillation has a less conclusive

pattern of DF distribution. Although some publications have shown a higher mean DF

in the left atria, particularly around the pulmonary veins (Wu et al. 2002), others have

shown that this distinction disappears for longstanding atrial fibrillation where other

34

sites are observed to predominate (Lazar et al. 2004; Atienza et al. 2006; Sanders et al.

2006; Stiles et al. online early). Isolation of the pulmonary veins abolishes this left‐to‐

right gradient in paroxysmal atrial fibrillation and, for the subset of patients with

persistent atrial fibrillation in whom a left‐to‐right gradient exists, success of

pulmonary vein isolation alone exceeds that of persistent atrial fibrillation without

such gradient (Lazar et al. 2006). Retrospective analysis of ablation of paroxysmal atrial

fibrillation shows that atrial fibrillation cycle length prolongation and eventual

termination occurs at sites of high DF (Sanders et al. 2005a). This has not been shown

for persistent atrial fibrillation where DF is distributed more evenly throughout atrial

regions, prompting a more aggressive multi‐region approach to be advocated when

ablating longstanding persistent atrial fibrillation (Haïssaguerre et al. 2005a). The

precise role of DF analysis to guide ablation clinically, like that of CFAE and

ganglionated plexi, is yet to be established.

1.3.6 The Neural Basis of Atrial Fibrillation

An additional dimension to the substrate of atrial fibrillation is that of the influence of

the autonomic nervous system. It has long been observed that electrical stimulation of

the autonomic nerves to the heart could induce atrial fibrillation (Lewis et al. 1921a;

Hoff and Geddes 1955). Atrial fibrillation has been proposed to be a consequence of

disordered autonomic tone (Coumel 1994). Neural influences on the atria mediated via

the autonomic ganglia found in the epicardial fat pads have been hypothesised to play

an important role in converting ectopic activity into atrial fibrillation (Scherlag et al.

2005a). This is based on the findings that during stimulation of canine atrial ganglia

35

fewer ectopic beats were required to initiate atrial fibrillation and, conversely, that

blockade of neural elements by lignocaine inhibits inducibility to atrial fibrillation by

ectopic beats. Applying this knowledge clinically, Scherlag et al. (2005b) added ablation

of the left atrial ganglionated plexi to the usual strategy of pulmonary vein antrum

isolation in 33 of 60 patients with paroxysmal or persistent atrial fibrillation. Those

receiving atrial denervation had less recurrence of atrial fibrillation at short‐term

follow up (median 5 months). Techniques by which to identify ganglionated plexi for

subsequent ablation have since been published. Using high‐frequency electrical

stimulation, an autonomic ganglion is identified by profound ventricular rate slowing

and subsequent ablation of the same area renders the stimulation response negative

on re‐testing (Lemery et al. 2006). Pappone et al. (2004) have published a 99% success

rate at 12 months follow up for patients who demonstrated complete vagal

denervation at their ablation procedure. Additionally, failure to achieve vagal

denervation was predictive for recurrence of atrial fibrillation. Takahashi et al. (2006)

characterised the change in atrial fibrillation cycle length occurring with vagally‐

mediated atrioventricular block during pulmonary vein isolation and found a greater

decrease in the cycle length of pulmonary vein activity than in the coronary sinus,

suggesting that vagal excitation enhances the driving role of pulmonary veins.

Recently, animal data linking the action of ganglionated plexi to complex fractionated

electrograms (Lin et al. 2007) and frequency gradient (Lu et al. 2008) have been

published. Convincing outcome data on the use of this technique to assist ablation is

yet to be published, but this remains a promising area of research.

36

1.3.7 Substrate Associated with Atrial Flutter

Experimental models of atrial flutter have shown abnormal electrical substrate to be

present. Morton et al. (2002) demonstrated in a sheep model that prolonged atrial

flutter resulted in abrupt‐onset reduction in refractoriness and slower‐onset reduction

in conduction velocity. Much work on atrial flutter has been performed in the sterile

pericarditis dog model which has been shown to dependably produce atrial flutter in a

time course similar to the post‐bypass arrhythmias seen clinically. The inducibility of

atrial flutter in this model is independent of differences in ERP, conduction time or

threshold (Page et al. 1986). More recently, the role of inflammation in atrial flutter

has been highlighted in a study in which the usually reliable induction of atrial flutter in

this model was abolished by pre‐treatment with prednisone (Goldstein et al. 2008).

Interestingly, this observation of no atrial flutter with steroid was accompanied by

lower ERP compared to the 11 untreated pericarditis dogs, strongly suggesting that

this aspect of electrical remodelling is neither integral to the initiation nor

maintenance of atrial flutter. Histological and gross inspection of the atria revealed

marked structural changes of inflammation in the untreated group, all of whom (bar

one with atrial fibrillation) developed atrial flutter. Therefore, suppression of

inflammation by corticosteroid is able to prevent subsequent arrhythmia in dogs given

sterile pericarditis which has intriguing implications for patients following cardiac

surgery. Patients with chronic atrial flutter have been observed to have electrical

remodelling by way of shortened monophasic action potential durations with impaired

rate adaptation compared to control patients (Franz et al. 1997). Sparks et al. (2000)

studied patients with paroxysmal atrial flutter and observed a reduction in ERP

immediately after 5‐10 minutes of induced atrial flutter, followed by a return to

37

baseline ERP over a period of minutes. In patients with chronic atrial flutter however,

ERP remained low out to 30 minutes and was seen to recover only when re‐tested at 3

weeks. These patients also demonstrated an improvement in corrected sinus node

recovery time (CSNRT) over the 3 week period suggesting sinus node dysfunction as a

direct result of chronic atrial flutter. Further evidence for remodelling of sinus node

function with atrial flutter is reported by Daoud et al. (2002) who studied patients with

chronic atrial flutter following ablation and observed impaired sinus node function

which recovered over a 3 month period. Differences in the underlying substrate seen

in atrial flutter between young and old have been published, with younger patients

tending toward slowest conduction in the lateral isthmus while older patients

demonstrate a low voltage zone and slowest conduction in the medial isthmus (Huang

et al. 2008). The substrate predisposing to atrial flutter may differ from that seen in

atrial fibrillation alone. Patients with typical atrial flutter in addition to paroxysmal

atrial fibrillation have been shown to have more frequent recurrence of arrhythmia

following pulmonary vein cryoisolation with accompanying isthmus line (Moreira et al.

2007). This suggests non‐pulmonary vein sites are more important to the maintenance

of atrial fibrillation in patients with co‐existent atrial flutter. This may reflect the

observation that an increase in the dispersion of refractoriness predicts the inducibility

of atrial fibrillation in patients undergoing ablation for atrial flutter (Ramanna et al.

2005).

38

1.3.8 Typical and Atypical Atrial Flutter

Electrophysiological studies have shown that the simple ECG definition of an organised

atrial rhythm with a rate typically between 250 and 350 beats per minute includes

many atrial tachycardias using a variety of reentry circuits. The reentry circuits often

occupy large areas of the atrium and are referred to as “macro‐reentrant” (Blomstrom‐

Lundqvist et al. 2003). Most early work on atrial flutter focussed on the more common

typical anti‐clockwise flutter; i.e. atrial flutter characterised by negative flutter waves

in the inferior ECG leads and a circuit moving in an anti‐clockwise direction when

viewed from the left anterior oblique radiographic view. This classic type of atrial

flutter is dependent on the cavotricuspid isthmus. Other forms of atrial flutter have

been described, including clockwise flutter and true atypical flutter. Clockwise atrial

flutter uses the same circuit with the same endocardial barriers as its anti‐clockwise

counterpart and is best considered a form of typical atrial flutter (Kalman et al. 1997).

It is amenable to the same ablation treatment as anti‐clockwise atrial flutter; i.e.

cavotricuspid isthmus ablation. Clockwise isthmus‐dependent atrial flutter shows the

opposite ECG pattern to the anti‐clockwise form – positive flutter waves in the inferior

leads and wide, negative flutter waves in lead V1, transitioning to positive waves in

lead V6. Olgin et al. (1997) showed that even in patients with clinical anti‐clockwise

atrial flutter, clockwise atrial flutter is frequently induced before ablation and is

dependent on the site of induction; pacing from the smooth right atrium induces anti‐

clockwise atrial flutter, whereas pacing from the trabeculated right atrium induces

clockwise atrial flutter. Atrial flutter caused by macro‐reentry circuits that do not use

the cavotricuspid isthmus are less common. Most are related to an atrial scar that

creates conduction block and a central obstacle for reentry (Blomstrom‐Lundqvist et

39

al. 2003). Prior cardiac surgery involving the atrium (such as repair of congenital heart

disease, mitral valve surgery or the atrial maze procedure) is a common cause

(Nakagawa et al. 2001; Triedman et al. 2001), as is (more latterly) ablation for atrial

fibrillation. True atypical atrial flutters usually show heterogeneous ECG morphologies,

shorter cycle lengths than typical atrial flutter, and frequent transitions from and to

atrial fibrillation (Kalman et al. 1997). Ablation of these macro‐reentrant loops is

particularly challenging as each circuit is unique to each patient. Furthermore, patients

with substantial atrial scarring may not demonstrate the ECG features of typical flutter

when cavotricuspid isthmus‐dependent flutter is indeed present, and conversely an

atypical flutter circuit may manifest on the surface ECG as common‐type atrial flutter

(Chugh et al. 2006).

1.4 Inter‐relationships between Atrial Flutter and Atrial Fibrillation

Atrial flutter has a close inter‐relationship with atrial fibrillation; the two rhythms

frequently co‐exist (Halligan et al. 2004) and share many of the same aetiological

factors (Vidaillet et al. 2002). The inter‐relationship is further distorted, in that for

some epidemiological reports the two conditions are reported together (Australian

Institute of Health and Welfare 2008). The substrate of atrial fibrillation however has

been more extensively investigated than that of atrial flutter and whether these two

arrhythmias share this substrate is unknown.

40

1.4.1 Atrial Flutter begets Atrial Fibrillation

The incidence of atrial fibrillation episodes prior to atrial flutter ablation is, perhaps

not surprisingly, a strong predictor for the occurrence of atrial fibrillation after the

procedure (Da Costa et al. 2002; Hsieh et al. 2002; Ramanna et al. 2005). However,

ablation of atrial flutter has been observed to significantly reduce the occurrence of

atrial fibrillation at 496±335 days post‐ablation from 55% to 33% (Schmieder et al.

2003). Despite this reduction a significant proportion, perhaps even the majority with

long‐term follow‐up, eventually develop atrial fibrillation (Chinitz et al. 2007; Meissner

et al. 2007; Moreira et al. 2008). Paydak et al. (1998) reported that 20±9 months after

atrial flutter ablation, 25% of patients have atrial fibrillation; Hsieh et al. (2002)

published a series of 333 patients in which 31% of patients had occurrence of atrial

fibrillation at 29±17 months follow up; and Ellis et al. (2007) described new atrial

fibrillation in 82% at 39±11 months follow up. Independent factors predictive of atrial

fibrillation within 6 months following ablation for atrial flutter are left ventricular

ejection fraction and, for those without prior atrial fibrillation, degree of mitral

regurgitation (Paydak et al. 1998; Da Costa et al. 2002). Yang et al. (2003) reported

mechanisms of transition from flutter to fibrillation observed in the laboratory during

atrial flutter ablation procedures as lower or upper loop reentry, and focal ectopic

activity (including pulmonary vein discharge). After 4 weeks of experimental atrial

flutter, conversion to atrial fibrillation has been demonstrated to be dependent on

critically timed extra‐stimuli, as demonstrated by Morton et al. (2002) in a sheep

model. It is being increasingly recognised that macro‐reentry may be a significant

contributor to the maintenance of long‐standing persistent atrial fibrillation

(Haïssaguerre et al. 2005b). Indeed, ablation of conduction gaps in the crista terminalis

41

can result in atrial fibrillation organising to atrial flutter (Liu et al. 2002). In the

subgroup of patients with atrial fibrillation in whom the administration of a class 1C

anti‐arrhythmic agent converts them to atrial flutter, it has been shown that atrial

flutter ablation, together with continuation of the drug, results in satisfactory control

of both arrhythmias (Schumacher et al. 1999).

1.4.2 Atrial Fibrillation begets Atrial Flutter

Key to the understanding of the close inter‐relationship of atrial flutter and atrial

fibrillation is the recognition that typical atrial flutter almost always develops from

antecedent atrial fibrillation of variable duration. The classic model of reentry seen in

clinical practice is that of atrioventricular reentry tachycardia. In this reentry circuit, a

critically timed extra‐stimulus encounters block in one limb of the circuit which is then

able to subsequently conduct in the opposite direction, thereby setting up a circuit. In

atrial flutter this pre‐existing circuit may not exist, therefore more than a simple

ectopic is required for its initiation. The line of functional inter‐caval block needs to be

created by a burst of rapid atrial rhythm for stable atrial flutter to ensue (Shimizu et al.

1991). Some patients have a fixed inter‐caval line of block and others have been shown

to have at least some degree of functional block (Olgin et al. 1995), therefore atrial

flutter may develop without preceding atrial fibrillation only in those with fixed block.

Waldo and Cooper (1996) observed triggers in post‐bypass patients that led to

transitional atrial fibrillation and then atrial flutter. Roithinger et al. (1997) observed

regular right atrial activation during atrial fibrillation which slowed and organised into

atrial flutter. If the pre‐requisite inter‐caval functional block is not achieved however,

42

either the atrial fibrillation will spontaneously revert to sinus rhythm or it will persist as

fibrillation rather than become atrial flutter. In short, there is no atrial flutter if the line

of functional block between the venae cavae does not form. It would therefore seem

that atrial fibrillation of variable duration (very brief to prolonged episodes) precedes

the onset of atrial flutter. The best evidence of this association is perhaps the study by

Ortiz et al. (1994) in the sterile pericarditis model on the association between the

length of inter‐caval block and atrial rhythm. During all episodes of stable pacing‐

induced atrial flutter, a line of functional block with a mean length of 24mm was

localised on the right atrial free wall. When the previously stable line of functional

block decreased to a mean of 16mm the line of functional block was not long enough

to maintain stable atrial flutter and conversion to atrial fibrillation resulted. When

sustained atrial fibrillation evolved to stable atrial flutter, there was reformation of a

long line of functional block, long enough (≥ prior length) to create a stable reentrant

circuit. Thus, the atrial rhythm was shown to depend directly on the length of

functional block between the inferior and superior venae cavae. Testing the hypothesis

that atrial fibrillation facilitates development of the line of functional conduction block

along the crista terminalis required to sustain atrial flutter, the effect of trigger

elimination by performing pulmonary vein isolation without cavotricuspid isthmus

ablation was studied by Wazni et al. (2003). In 59 patients with both atrial fibrillation

and atrial flutter, 32 of these patients developed episodes of atrial flutter acutely.

However, only 3 had sustained atrial flutter after 8 weeks suggesting that, without the

antecedent atrial fibrillation, atrial flutter was unable to be initiated. This concept of

treating fibrillation to cure flutter still requires further evaluation before becoming

mainstream therapy. Thus, it can be seen that atrial flutter and atrial fibrillation are

43

closely inter‐related and that the features previously shown in the underlying atrial

substrate of atrial fibrillation are likely to be also present in atrial flutter, even in those

in whom no clinical history of atrial fibrillation is noted. Critical to our management of

patients with both atrial flutter and atrial fibrillation is an understanding of the

interactions and shared characteristics of these rhythms.

1.5 Rate‐related Remodelling

1.5.1 Electrical Remodelling

Wijffels et al. (1995) observed that in normal goats, induction of atrial fibrillation was

relatively difficult and lasted for just a few seconds. With repetitive induction of

arrhythmia however, the cycle length of the atrial fibrillation shortened and episodes

became sustained. This led to the notion that “atrial fibrillation begets atrial

fibrillation” by means of shortening in atrial ERP after periods of atrial fibrillation.

Further experimental studies have confirmed this finding in dogs (Morillo et al. 1995;

Elvan et al. 1996; Goette et al. 1996) with the additional observations that atrial

fibrillation caused an increase in the heterogeneity of refractoriness (Fareh et al. 1998)

and maladaption of refractoriness to rate (Lee et al. 1999). While initial studies did not

observe conduction abnormalities, subsequent canine studies by Gaspo et al. (1997)

demonstrated that progressively longer episodes of atrial fibrillation also resulted in

conduction slowing. Attenuation of conduction velocity was maximal at 6 weeks, and

lagged behind that of maximum reduction in ERP at 7 days, suggesting that these

abnormalities are mediated by different mechanisms. Morton et al. (2002)

demonstrated a reduction in atrial refractoriness in a sheep model with sustained

44

atrial flutter. Again a reduction in conduction velocity was seen to develop over a

slower time course. Thus, the mechanism by which electrical remodelling of the atria

lends toward the perpetuation of arrhythmia in response to atrial fibrillation is by a

shortening of the reentrant “wavelength”. Human studies of patients with recent atrial

arrhythmia have also been shown to have shorter ERP than control groups, as well as

conduction delay and sinus node dysfunction (Cosio et al. 1983; Kumagai et al. 1991).

Yu et al. (1999) demonstrated shorter ERP with impaired rate adaptation and

depressed conduction following cardioversion for longstanding atrial fibrillation, with

the ERP gradually increasing after 4 days of sinus rhythm. Garratt et al. (1999)

evaluated the effect of repeated episodes of atrial fibrillation on atrial electrical

remodelling. While marked remodelling was observed with five days of atrial

fibrillation, two days of sinus rhythm was adequate to reverse all abnormalities.

Despite repetitive episodes of atrial fibrillation, there were no additive changes

observed in atrial electrical remodelling. Sparks et al. (2000) documented the changes

in atrial ERP following ablation for chronic atrial flutter and observed a recovery in ERP

over a 3 week period. In the same study, 10 minutes of induced atrial flutter in a group

of patients presenting in sinus rhythm for their flutter ablation reduced ERP but this

recovered within 15 minutes. To investigate the possibility that early and repeated

cardioversion of atrial fibrillation could avoid the cycle of adverse electrical

remodelling, Fynn et al. (2002) evaluated this strategy in the clinical setting. Despite an

aggressive protocol to achieve and maintain sinus rhythm, frequent cardioversion

failed to prevent the progression to more atrial fibrillation. This occurred despite a well

documented increase in atrial ERP at successive cardioversions indicating that reversal

of electrical remodelling was in fact taking place. Indeed, their findings suggest that in

45

patients with a history of atrial fibrillation, sinus rhythm does not beget sinus rhythm.

In addition to the effects on ERP and conduction, several investigators have observed

remodelling of the sinus node that develops soon after the initiation of arrhythmia and

reverses after restoration of sinus rhythm. Kumagai et al. (1991) reported significantly

longer sino‐atrial conduction time (SACT) and CSNRT in patients recently cardioverted

from chronic lone atrial fibrillation compared to controls. Elvan et al. (1996)

demonstrated in dogs that atrial fibrillation was responsible for impaired sinus node

function which partially reversed when sinus rhythm was restored. Other investigators

have confirmed the prolongation of CSNRT in patients following cardioversion for atrial

fibrillation compared to control patients (Tse et al. 1999) and furthermore this finding

may be associated with a trend to recurrence of atrial fibrillation (Manios et al. 2001).

This phenomenon of sinus node remodelling does not appear unique to atrial

fibrillation but has also been demonstrated with atrial flutter (Sparks et al. 2000;

Daoud et al. 2002). Hocini et al. (2003) demonstrated in 20 patients with a

documented >3 second pause post‐termination of atrial fibrillation that isolation of the

pulmonary veins and linear atrial ablation improved sinus node function as measured

by ambulatory heart rate parameters and CSNRT at 2 years. 19 of these patients

avoided the implantation of a pacemaker, including 1 with ongoing infrequent

episodes of atrial fibrillation.

1.5.2 Structural Remodelling

While the above electrical changes observed due to atrial arrhythmia are a relatively

recent finding, Davies and Pomerance (1972) were the first to suggest structural

46

change due to atrial fibrillation itself. After post‐mortem examination of 100 patients

with atrial fibrillation, they described loss of atrial myocardium in the region of the

sinus node and nodal artery stenosis in patients with long‐term atrial fibrillation, as

compared to the lack of such features in patients in whom arrhythmia had only been

present within the last 2 weeks of their life. Additional data from tissue of patients

taken at cardiac bypass operations reported degenerative changes in the atria of

patients with arrhythmias (Mary‐Rabine et al. 1983). Further characterisation of atrial

tissue from patients undergoing cardiac surgery has revealed apoptotic death of

myocytes in those with atrial fibrillation, suggesting some degree of irreversible

remodelling (Aimé‐Sempé et al. 1999). Other changes at the cellular level have been

observed in the distribution of connexin sub‐types in patients with atrial fibrillation

(Kostin et al. 2002). In patients with sinus rhythm, gap junction proteins are localised

to the intercalated discs, whereas in patients with atrial fibrillation there is

lateralisation of these proteins and a preferential expression of connexin 40 over

connexin 43. Ausma et al. (2003) have observed that ultrastructural changes evolve

over a much longer time frame than the electrical remodelling of the atria in response

to rate, and alterations in ultrastructure are not completely resolved as late as four

months after return to sinus rhythm from prolonged atrial fibrillation. These slowly

resolving alterations have been characterised as an increase in myolysis and the

fraction of extracellular matrix per myocyte, as well as changes in the expression of

structural proteins. It is these structural changes associated with atrial fibrillation

occurring over a slower time‐course than that of electrical remodelling which have

been proposed as the critical factors in the “domestication” of atrial fibrillation

(Allessie et al. 2002).

47

1.6 Atrial Substrate in Conditions Associated with Atrial Arrhythmia

As discussed in Sections 1.1.2 and 1.1.4, several clinical conditions have been shown to

have an association with atrial fibrillation and atrial flutter. Study of the atrial substrate

in conditions predisposing to atrial arrhythmia may give insights into the factors

promoting rhythm disturbance.

1.6.1 Heart Failure

In patients with symptomatic heart failure, the prevalence of atrial fibrillation ranges

from 10 to 30 percent, with the highest incidence among those with the most severe

heart failure (Stevenson and Stevenson 1999). The electrophysiological substrate of

the atria present in heart failure has been studied extensively. Boyden et al. (1984)

demonstrated large amounts of interstitial fibrosis, cellular hypertrophy and

degeneration, and thickened basement membranes in the atria of a feline model of

cardiomyopathy. Power et al. (1998) showed in a ventricular pacing‐induced sheep

model of heart failure, that larger left atria were increasingly susceptible to atrial

fibrillation. Li et al. (1999) published findings from a canine model of heart failure that

showed an increase in the heterogeneity of conduction associated with histological

changes of marked interstitial fibrosis, and furthermore that these changes are

mediated via significant alterations in cellular electrophysiology (Li et al. 2000). Clinical

studies of heart failure have demonstrated that electrophysiological and structural

remodelling is present even in a cohort without known atrial fibrillation (Sanders et al.

2003). Studies of the right atrium showed that voltage reduction and areas of electrical

silence were more apparent in the heart failure group when compared to a control

48

group, and that atrial refractoriness was increased in the diseased atria. Conduction

delay was recorded not only in global measures and regional analysis, but also

specifically at the crista terminalis in the heart failure group who were more

susceptible to atrial fibrillation.

1.6.2 Atrial Septal Defects

Patients with haemodynamically‐significant atrial septal defects are characterised by

chronic right atrial volume overload and are at increased risk of atrial arrhythmia;

furthermore this risk persists despite correction of the underlying anatomical defect

(Murphy et al. 1990). Morton et al. (2003), in studying 13 patients with atrial septal

defects, demonstrated an increase in right atrial refractoriness, conduction delay at

the crista terminalis and sinus node dysfunction compared to a control group.

Importantly, re‐study in 4 of these patients at >6 months following correction of the

defect showed that site‐specific conduction delay persisted while there was no

significant change in the electrical parameters, perhaps explaining why risk of

arrhythmia has been shown to persist.

1.6.3 Hypertension

Hypertension is the most common contributor to the burden of atrial fibrillation in the

developed world, simply due to its high prevalence in Western society (Kannel et al.

1982a). Increases in left atrial dimension and left ventricular mass as well as

prolongation of the maximum duration and dispersion of the P wave, and reduced A‐

wave velocity are predictors for the onset of atrial fibrillation in the hypertensive

49

population (Ciaroni et al. 2000). Kistler et al. (2006) demonstrated in an ovine model of

hypertension that widespread conduction abnormalities and shortening of atrial

wavelength were present when compared to controls. The increase in atrial fibrillation

that they observed occurred despite no significant change in refractoriness. Atrial

myocyte hypertrophy and myolysis were seen in all hypertensive sheep and focal

scarring in many. Mitochondrial and nuclear enlargement was seen under electron

microscopy in hypertensive sheep but not in any controls.

1.6.4 Valvular Heart Disease

Mitral regurgitation is associated both with atrial fibrillation and congestive heart

failure; furthermore onset of arrhythmia often heralds haemodynamic

decompensation (Grigioni et al. 2002). Animal models of mitral incompetence have

shown that while conduction is unchanged after chronic atrial volume overload, a

uniform increase in atrial refractoriness is observed suggesting that electrical

remodelling is not the means by which atrial fibrillation is inclined to develop in this

disease (Verheule et al. 2003). Similarly, mitral stenosis predisposes to atrial

fibrillation. Boyden et al. (1982) studied 23 dogs with spontaneous mitral valve fibrosis

and left atrial enlargement. Although these animals demonstrated no difference in

action potential duration it was noted that there were reduced numbers of muscle cell

layers and an unusually large amount of connective tissue between greatly

hypertrophied cells. Atrial fibrillation is a disease with which rheumatic mitral stenosis

is not only frequently associated, but is in fact a potentially disastrous complication of,

leading to a 17‐fold increase in the risk of thromboembolism (Wolf et al. 1978). Fan et

50

al. (2002) performed electrophysiological evaluation of the right atrium in 31 patients

at the time of mitral commissurotomy for rheumatic mitral stenosis. Following

cardioversion of atrial fibrillation, these patients demonstrated shorter ERP, sinus node

dysfunction and no difference in atrial conduction compared to those patients who

presented for the procedure in sinus rhythm. Soylu et al. (2004) evaluated right atrial

ERP before and after mitral commissurotomy in 25 patients with mitral stenosis in

sinus rhythm. This study observed an acute increase in ERP immediately after relief of

mitral stenosis. However, neither of these studies included a normal control group to

determine the effects of mitral stenosis independent of the effects of the arrhythmia

itself or provided left atrial data. Recently, John et al. (online early) have described in

detail the electrophysiological substrate of both the right and left atria of a group with

severe mitral stenosis undergoing mitral valvuloplasty. This study demonstrated

widespread and severe atrial remodelling when compared to a control group with left

sided accessory pathways, despite not having previously had atrial fibrillation. This

suggests that it was the structural remodelling, in this case characterised by voltage

reduction, conduction slowing and sinus node dysfunction, which predisposed this

group to their increased susceptibility to atrial fibrillation during ERP testing. Indeed,

the ERP testing revealed an increase in atrial refractoriness which runs counter to the

electrical remodelling seen following episodes of atrial fibrillation (Wijffels et al. 1995)

and is more in keeping with the “atrial remodelling of a different sort” first reported by

Li et al. (1999) in another condition predisposed to atrial fibrillation.

51

1.6.5 Sub‐clinical Atrial Disease in Lone Atrial Fibrillation

Frustaci et al. (1997) reported abnormal atrial histology including inflammatory

infiltrates and patchy fibrosis in all of 12 patients studied with paroxysmal lone atrial

fibrillation, and in none of 11 control patients. Excess fibrosis has been characterised

as an increased concentration of collagen type I in patients with lone atrial fibrillation

compared to controls and this increase has been observed to be attenuated in those

patients receiving angiotensin converting enzyme inhibitors (Boldt et al. 2006). Further

evidence that inflammation plays a role in atrial fibrillation comes from the

observation that C‐reactive protein is elevated in patients with lone atrial fibrillation

(Chung et al. 2001). Diastolic dysfunction diagnosed by transeptal haemodynamic

evaluation of left atria has been demonstrated to be present in patients with lone

atrial fibrillation (Jaïs et al. 2000). In addition, there is some evidence to suggest that

patients with lone atrial fibrillation have microvascular dysfunction in the coronary

circulation (Skalidis et al. 2008). Thus, it is feasible that patients without overt

structural heart disease may have fibrosis, inflammation, structural change and

vascular dysfunction that contribute to the underlying atrial substrate. In addition,

when considering the atrial substrate encountered in diseases known to be associated

with the subsequent development of atrial arrhythmia some unifying features can be

seen to emerge. Atrial refractoriness is either unchanged or increased, while structural

abnormalities, conduction delay and sinus node dysfunction become increasingly

apparent. Cellular and ultrastructural alteration underpins these gross and functional

changes and the end result of such abnormal substrate is an increase in the likelihood

of initiation and maintenance of atrial arrhythmia.

52

1.7 Sinus Node Disease, Remodelling and Atrial Arrhythmias

“The sinus node is the essence of life” – Bruno Kisch, co‐founder and president of the

American College of Cardiology, 1951‐1953.

Remodelling of the sinus node is known to develop soon after the initiation of

arrhythmia and reverse after restoration of sinus rhythm (see Section 1.5.1).

Irrespective of the mechanism by which sinus node remodelling occurs due to rapid

atrial rates, an increase in the time window due to sinus bradycardia can facilitate the

development of atrial fibrillation by increasing atrial ectopy and the dispersion of

refractoriness (Luck and Engel 1979). Therefore, the frequent co‐existence of atrial

arrhythmia and sinus node disease may be due to a slow sinus rate increasing

vulnerability to atrial fibrillation, a common disease substrate, and/or arrhythmia

directly impairing sinus node function; all 3 possibilities have evidence to support them

(Rubenstein et al. 1972; Moss and Davis 1974; Gomes et al. 1981).

1.7.1 The Anatomical Sinus Node

The anatomical sinus node is located at the junction of the superior vena cava with the

right atrium (James 1977; Anderson et al. 1979). The sinus node consists of specialised

nodal cells lying immediately sub‐epicardially at the superior pole of the sulcus

terminalis of the right atrium which, under autonomic influence, spontaneously

depolarise at a variable rate to match heart rate to physiological demand. The

endocardial aspect of the sulcus terminalis is marked by the crista terminalis; the

structure demarcating the embryonic junction of the anterior trabeculated appendage

and the posterior smooth‐walled venous component. In a study of 47 adult hearts, the

53

histological sinus node has been shown to be crescent‐shaped with the long axis

parallel to the crista terminalis and has a mean length of 13.5mm (range 8‐21.5mm)

(Sanchez‐Quintana et al. 2005). Histological sectioning of the atria by a range of

investigators has failed to identify islands of nodal cells separate from the sinus node

itself. Indeed, all such studies have demonstrated that the sinus node has a defined

and consistent anatomical location (James 1977; Anderson et al. 1979; Boineau et al.

1980; Sanchez‐Quintana et al. 2005). No insulating tract is seen and nodal tissue

radiates for short distances into the surrounding working myocardium. In comparison

to contractile cardiac myocytes, nodal cells are smaller cells that are packed within a

dense matrix of connective tissue. The lack of intercalated discs and poorly developed

sarcoplasmic reticulum are proposed reasons for slow intra‐nodal conduction (James

et al. 1966). Recent immunohistochemical studies on the sinus node have revealed

regional differences in connexin expression. Kwong et al. (1998) showed that the

canine sinus node was composed of mainly connexin 43‐negative cells, with the

majority expressing connexin 40. The one‐third of cells expressing connexin 43

however, abutted working atrial myocardium and it is hypothesised that these cells

form bridges for the propagation of the action potential from the centre of the sinus

node to the atrial muscle.

1.7.2 The Functional Sinus Node Complex

Shifts in the origin of the sinus node impulse along the sulcus terminalis under

autonomic influences were first observed by Lewis et al. (1910) and by Meek and

Eyster (1914) early last century. The advent of medium‐density epicardial mapping

54

using multi‐polar epicardial plaques confirmed the initial observations, with the

functional sinus node demonstrating complex activity under autonomic control

(Bouman et al. 1968; Geesbreght and Randall 1971; Goldberg 1975; Goldberg et al.

1981). Boineau and colleagues (1978; 1980) performed medium‐density mapping of

the right atrial epicardium with simultaneous endocardial mapping in a subset,

together with some histological analysis, under conditions of sympathetic and

parasympathetic stimulation and following overdrive pacing. They observed that the

sinus P wave may arise from a ‘pacemaker complex’, distributed over an extensive

extra‐nodal area spanning the superior vena cava ‐ right atrial junction to the inferior

vena cava. These studies of the canine functional sinus node complex have identified

that the area of the origin of cardiac depolarisation exceeds that of the anatomical

sinus node area by 3 to 4‐fold. A strong correlation between heart rate and functional

sinus origin was observed with vagal stimulation and isoprenaline infusion resulting in

inferior and supero‐anterior shifts, respectively. This study reported several seminal

observations; (i) at least three anatomically fixed (in contrast to diffusely wandering)

areas of sinus node origin were present, each of which became dominant within a

specific range of heart rates, (ii) after intramural disconnection of the origins they

maintained their normal function in their respective heart rate ranges, and (iii) origin

points were mapped to the same anatomical location using both epicardial and

endocardial mapping, which suggested a lack of intramural preferential conduction.

However, in conflict with their proposition was the histological absence of specialised

nodal cells at the extra‐nodal sinus origins. On the weight of evidence, Boineau et al.

(1980) concluded that the multi‐centric nature of functional sinus origin was mediated

through a complex system of separate atrial pacemakers with differential sensitivities

55

to autonomic control. Boineau et al. (1988) extended their studies to humans to

elegantly describe the functional sinus node origin(s) using 156 bipole epicardial

plaques in 14 patients undergoing open‐heart surgery for the ablation of accessory

pathways. They reported uni‐focal nodal and extra‐nodal sinus origins in addition to

multi‐centric origins over two to four widely distributed pacemaker sites. The similarity

in anatomical origin of escape beats after overdrive pacing compared to that in sinus

rhythm supported the role of extra‐nodal sites in healthy sinus function. The functional

sinus pacemaker complex in humans was estimated to exist over a zone centred about

the long axis of the sulcus terminalis encompassing a 7.5 by 1.5cm area, with a cranial

and caudal border consisting of the superior and inferior venae cavae, respectively.

Alternative hypotheses for extra‐nodal sinus origins proposed by Boineau et al. (1988)

include; (i) the existence of specialised conduction tracts from the anatomically fixed

nodal source to remote exit sites, and (ii) a coordinated exchange of dominance

amongst competing pacemakers via autonomic modulation. For the former

explanation, the presence of nodal radiations into adjacent atrial myocardium through

electrically isolating connective tissue supports this theory, however no insulated

rapidly conducting tissue analogous to that found in the ventricles has been

conclusively demonstrated (Anderson et al. 1981a). Supporting the latter hypothesis,

autonomically mediated shifts in activation sites appear linked to changes in heart

rate, however the mechanism by which a stable rate is maintained without faster

pacemakers dominating is yet to be shown. These considerations led to a hybrid model

being proposed where neural and hormonal factors influence both the site of

pacemaker activation and the point of exit from the sinus node (Schuessler et al.

1996). Additional canine right atrial studies have shown that transmembrane

56

potentials from sinus node pacemaker cells do not show a temporal or spatial

relationship with atrial extracellular potentials (Bromberg et al. 1995). This suggests

relative electrical isolation of the sinus node from surrounding working atrial

myocardium with the earliest extracellular potentials representing the exit points from

the node. Further evidence of a multi‐centric and widely distributed functional

pacemaker complex comes from clinical studies that have aimed to modify sinus

activity in patients with inappropriate sinus tachycardia. Inappropriate sinus

tachycardia is a non‐paroxysmal tachyarrhythmia characterised by an increased resting

heart rate and/or an exaggerated heart rate response to minimal exertion or a change

in body posture (Still et al. 2005). In contrast to a single uni‐focal lesion, procedures for

the treatment of inappropriate sinus tachycardia require extensive ablation along the

entire length of the crista terminalis to achieve functional inhibition of the sinus node

(Kalman et al. 1995; Lee et al. 1995).

1.7.3 Conventional Evaluation of Sinus Node Function

Conventional tests of sinus node automaticity and sino‐atrial coupling are based on

indirect surrogate measures and as such, are thought to contain limited clinical value

(Tonkin and Heddle 1984). Sinus node automaticity has traditionally been assessed by

the duration of the first return cycle after cessation of sinus overdrive pacing (Mandel

et al. 1971; Narula et al. 1972). The CSNRT adjusts for the preceding sinus cycle length

before overdrive pacing, however neither the raw measure nor CSNRT provide a pure

assessment of sinus node automaticity and both are confounded by extrinsic

influences. Consequently, there is poor inter‐laboratory agreement for ‘normal’ ranges

57

of CSNRT (Kulbertus et al. 1975; Breithardt et al. 1977; Alboni et al. 1982) and the

positive predictive accuracy in patients with suspected sinus node dysfunction is as low

as 30% (Yee and Strauss 1987). Methods to evaluate sino‐atrial conduction have

included indirect measures (Strauss et al. 1973; Narula et al. 1978) and the direct

method by the recording of a sinus node electrogram (Hariman et al. 1980). However,

as with measures of automaticity, the positive predictive value of these tests for

identification of clinical sinus node disease has been too low to warrant routine clinical

use. Nevertheless, due to their widespread use within the body of literature on sinus

node disease, these measures retain a degree of utility for research on the mechanism

of sinus node dysfunction.

1.7.4 High‐density Mapping of the Sinus Node

Initial mapping studies of the functional sinus node complex utilised multi‐electrode

plaques placed epicardially or endocardially at surgery (Boineau et al. 1978). The

subsequent development of 3‐dimensional endocardial mapping systems has allowed

less invasive assessment of sinus node activity, although high‐density has been

achieved using analysis of many sequential beats rather than single‐beat mapping

(Sanders et al. 2004b). The advent of non‐contact mapping has facilitated high‐density

simultaneous endocardial mapping of cardiac depolarisation. The morphology of

unipolar electrograms provides measurable information of the depolarisation wave‐

front direction, speed and width of the approaching or retreating wave. In an animal

model, virtual unipolar electrograms from non‐contact mapping were able to

differentiate between endocardial and sub‐endocardial pacing origins thus providing

58

information on the intramural depth of impulse origin (Kroll et al. 2003). The

morphology characteristics of virtual electrograms have been shown to match with

high reliability those of conventional contact unipolar electrograms (Schilling et al.

1998; Kadish et al. 1999; Schilling et al. 2000; Thiagalingam et al. 2004a; Earley et al.

2006). Higa et al. (2004) have shown insights into the additional information provided

via the utilisation of high density unipolar mapping during atrial tachycardia. Non‐

contact mapping showed that in many cases the traditionally mapped origin – the site

from which centrifugal activation occurs to spread activation to the remaining atria –

was indeed a break‐out point. Detailed inspection of the unipolar electrograms

demonstrated an earlier site of origin (“QS” morphology) with a preferential pathway

of conduction to the break‐out point. These investigators were able to confirm the

existence of an alternate origin with a preferential conduction pathway by curative

ablation of the tachycardia at both the origin and also along the preferential pathway.

Whether similar pathways exist in the non‐diseased atrium, including within the sinus

node, is yet to be reported.

1.7.5 Sinus Node Disease

Sinus node disease is thought to result from disordered impulse generation within the

sinus node or impaired conduction of the impulse to the surrounding atrial tissue

(Ferrer 1968). While differentiation between these two potential aetiologies has been

difficult owing to the absence of an appropriate experimental model, Sanders et al.

(2004a) have performed point‐by‐point atrial mapping studies in patients with sinus

node disease and found evidence of a diffuse atrial myopathy. Compared to age‐

59

matched controls, patients with sinus node disease demonstrated an increase in atrial

refractoriness, atrial conduction time and P wave duration. Interestingly, the sinus

node complex in disease was more often uni‐centric and localised to the low crista

terminalis with evidence of regional conduction slowing and scar. These data suggest

that impulse conduction rather than impulse generation are responsible for the

manifestations of sinus node disease.

1.7.6 Clinical Conditions Associated with Sinus Node Impairment

An association with sinus node dysfunction is seen for many disease states. This not

only includes the direct relationship between atrial arrhythmia and sinus node

remodelling, but also conditions (e.g. heart failure) that demonstrate an association

with both atrial arrhythmia and impairment of sinus node function. Elvan et al. (1996)

demonstrated significant prolongation of CSNRT and the intrinsic sinus cycle length

after two to six weeks of atrial fibrillation in dogs. In humans, short durations of rapid

atrial pacing resulted in similar sinus node inhibition (Hadian et al. 2002) and CSNRT

was significantly longer in patients cardioverted from chronic lone atrial fibrillation

compared to controls (Kumagai et al. 1991; Tse et al. 1999; Manios et al. 2001).

Manios et al. (2001) also reported that sinus node inhibition improved within 24 hours

after cardioversion. Sinus node dysfunction as measured by traditional parameters has

been reported after reversion to sinus rhythm from atrial flutter (Sparks et al. 2000;

Daoud et al. 2002). A decrement in sinus node function has been reported during

clinically induced atrial stretch. In a prospective randomised comparison, Sparks et al.

(1999) reported that long‐term (3 months) VVI pacing resulted in increased atrial

60

refractoriness, a prolongation of P wave duration and longer CSNRT in comparison to

DDD pacing. They further highlighted that the changes observed in the VVI pacing

group were reversed with three months of DDD pacing. Similar increases in P wave

duration and CSNRT have been reported in patients with atrial stretch due to

significant atrial septal defects (Morton et al. 2003) and congestive heart failure

(Sanders et al. 2004b). For the latter, Sanders et al. reported a more caudal localisation

of sinus origin and greater structural remodelling of the crista terminalis in heart

failure patients when compared to age‐matched controls. Structural remodelling was

represented by a loss of voltage, demonstration of conduction delay at the crista

terminalis and circuitous propagation of the sinus impulse. These findings raise the

possibility that dysfunction of the sinus node may have been a consequence of

structural change associated with atrial stretch.

61

Chapter 2. Right and left atrial electrical substrate of lone atrial fibrillation

2.1 Introduction

Atrial fibrillation arises as a result of a complex interaction of triggers, perpetuators

and substrate (Allessie et al. 2001). Experimental studies of atrial fibrillation have

demonstrated shortening of the ERP and slowing of conduction as a result of atrial

fibrillation and have suggested that these factors combine to promote continuing atrial

fibrillation, giving rise to the concept that “atrial fibrillation begets atrial fibrillation”

(Morillo et al. 1995; Wijffels et al. 1995; Goette et al. 1996; Gaspo et al. 1997; Fareh et

al. 1998). Clinical studies have shown reversal of electrical remodelling over time after

termination of arrhythmia (Yu et al. 1999; Hobbs et al. 2000). However, strategies of

prompt termination of atrial fibrillation to avoid this cycle of adverse remodelling have

failed to show benefit (Fynn et al. 2002). In fact, the natural history of paroxysmal

atrial fibrillation is one of increasing frequency and duration of episodes (Kopecky et al.

1987). These observations have led to the search for a “second factor” integral to the

development and progression of atrial fibrillation (Garratt et al. 1999).

We hypothesised that patients with paroxysmal “lone” atrial fibrillation have an

abnormal atrial substrate that determines disease progression. In order to evaluate

the substrate predisposing to atrial fibrillation, free of the electrical remodelling effects

of the arrhythmia itself, we evaluated the electrophysiological characteristics of atria

in patients with paroxysmal “lone” atrial fibrillation, remote in time from the

arrhythmia.

62

2.2 Methods

2.2.1 Study Population

This study comprised 13 patients undergoing first‐time ablation for paroxysmal atrial

fibrillation and a reference group of 13 patients with structurally normal hearts

undergoing radiofrequency ablation for atrioventricular reentry tachycardia with left‐

sided accessory pathways and no history of atrial fibrillation. Patients for the study

group were excluded if they had atrial fibrillation during the week prior to ablation

(established by continuous monitoring) or any of the criteria that would prohibit the

diagnosis of “lone” atrial fibrillation, previously defined as the absence of structural


blood chemistry, and trans‐thoracic as well as trans‐oesophageal echocardiography

(Kopecky et al. 1987; Kumagai et al. 1991; Jaïs et al. 2000). Coronary artery disease was

excluded by clinical, ECG or stress test criteria. Pulmonary disease, hypertension,

hyperthyroidism and diabetes were eliminated by appropriate tests. Paroxysmal atrial

fibrillation was defined according to the Expert Consensus Statement as recurrent

atrial fibrillation that terminates spontaneously within 7 days (Calkins et al. 2007).

Thus, of 109 patients with highly symptomatic atrial fibrillation presenting for ablation

screened, 13 patients were suitable to be included in this study.

All anti‐arrhythmic medication was ceased ≥5 half‐lives prior to the study. No patient

had received amiodarone. All patients provided written informed consent to the study

protocol which was approved by the Human Research Ethics Committee of the Royal

Adelaide Hospital.

63

2.2.2 Electrophysiological Study and Ablation

Electrophysiological study was performed in the post‐absorptive state with sedation

utilising midazolam and fentanyl. In patients with atrial fibrillation the study protocol

was performed before ablation, while in the reference group this was done after

accessory pathway ablation. The left atrium was accessed using a single transeptal

puncture following which repeated bolus unfractionated heparin was utilised to

maintain the activated clotting time between 300‐350 seconds.

Following the study protocol, patients with atrial fibrillation underwent circumferential

pulmonary vein ablation with an end‐point of isolation confirmed by circumferential

mapping (Lasso, Biosense‐Webster) with either elimination or dissociation of

pulmonary venous potentials. Ablation of the pulmonary veins was performed using a

delivered power of 30W with irrigation rates of 30mL/min (Thermocool, Biosense

Webster). Additional substrate modification was performed in patients with an

episode of atrial fibrillation ≥48 hours or with a left atrial size ≥57mm (longest

diameter). This took the form of linear ablation along the left atrial roof with an end‐

point of linear double potentials and conduction block demonstrated by an activation

detour during pacing of the left atrial appendage. Cavotricuspid isthmus ablation with

an endpoint of bi‐directional isthmus block was performed only in patients with a

history of typical flutter or if mapping confirmed typical flutter during the procedure.

Linear ablation was performed with a delivered power of 30‐35W with irrigation rates

of 30‐60mL/min.

64

2.2.3 Electrophysiology Study Protocol

The following catheters were positioned for the study protocol: (i) a 10‐pole catheter

(2‐5‐2mm inter‐electrode spacing, Daig Electrophysiology) within the coronary sinus

with the proximal bipole at the coronary sinus ostium as determined in the best septal

left anterior oblique position; (ii) a 20‐pole “crista” catheter (1‐3‐1mm inter‐electrode

spacing, Biosense‐Webster) placed along the crista terminalis with the distal tip

superiorly such that the second bipole lay at the junction of the superior vena cava and

right atrium, stabilised by a long sheath (CSTA, Daig Electrophysiology) to ensure close

apposition to this structure; (iii) a 20‐pole catheter (2‐5‐2mm inter‐electrode spacing,

Daig Electrophysiology) placed along the lateral right atrium; and (iv) a roving 10‐pole

catheter (2‐5‐2mm inter‐electrode spacing, Biosense‐Webster) was positioned within

the left atrium via trans‐septal puncture (Figure 2‐1). This catheter was stabilised with

the use of a long sheath (Preface, Biosense‐Webster or SL0, Daig Electrophysiology)

and sequentially positioned as follows at the: (i) left atrial roof; (ii) inferior left atrium;

(iii) mid‐posterior left atrium; (iv) left atrial appendage and (v) high right atrial septum,

as previously described (John et al. online early). Electrophysiological evaluation was

performed as detailed below.

2.2.3.1 Effective Refractoriness

Atrial ERP was evaluated at twice diastolic threshold at cycle lengths of 600 and 450ms

using an 8‐beat drive followed by an extra‐stimulus, starting with a coupling interval of

150ms increasing in 10ms increments. ERP was defined as the longest coupling interval

failing to propagate to the atrium. At each site the ERP was measured 3 times at each

cycle length and averaged. If ERP varied by >10ms an additional 2 measurements were

65

made and the total number averaged. ERP was measured from the following sites: (i)

distal coronary sinus; (ii) proximal coronary sinus; (iii) low‐lateral right atrium; (iv) high‐

lateral right atrium; (v) high‐septal right atrium; (vi) left atrial appendage; (vii) mid‐

posterior left atrium; (viii) inferior left atrium; (ix) junction of the left superior

pulmonary vein with the left atrial roof; and (x) junction of the right superior

pulmonary vein with the left atrial roof. Atrial fibrillation induced by ERP testing lasting

>5 minutes was considered sustained; when this occurred, no further data was

acquired.

2.2.3.2 Atrial Conduction

Atrial conduction time was assessed along linearly placed catheters by pacing the distal

bipole (1,2) and determining the conduction time to a proximal bipole (9,10) at the left

atrial roof, inferior left atrium, coronary sinus and lateral right atrium. Conduction time

at each site was averaged over 10 beats during stable capture at 600 and 450ms cycle

lengths.

P wave duration was determined as a marker of global conduction by the average of

10 beats on ECG lead II.

2.2.3.3 Site‐specific Conduction Abnormalities at the Crista Terminalis

The number of bipoles on the crista terminalis catheter with discrete double potentials

separated by an isoelectric interval or complex fractionated activity of ≥50ms duration,

and the maximum electrogram duration were determined during sinus rhythm,

constant pacing and for the shortest‐coupled captured extra‐stimulus from the

66

proximal coronary sinus, low‐lateral right atrium, inferior left atrium and left atrial

appendage.

2.2.3.4 Sinus Node Function

Sinus node function was evaluated as follows: (i) Baseline sinus cycle length was

determined over 10 consecutive sinus cycles; (ii) SACT was determined after an 8‐beat

pacing train [using the formula SACT = (return ‐ basic cycle length)/2] (Narula et al.

1978) 3 times and averaged; and (iii) CSNRT was determined after a 30‐second drive

train at cycle lengths of 600 and 450ms, correcting for the baseline cycle length. At

each cycle length, CSNRT was determined 3 times and averaged.

2.2.4 Statistical Analysis

Continuous variables are reported as mean ± standard deviation or median and inter‐

quartile range as appropriate. Categorical variables are reported as number and

percentage. Proportions were compared using Fisher’s Exact test. Comparisons

between means were analysed using paired or unpaired t tests as appropriate.

Comparisons with adjustment for multiple sampling within patients were performed

using a Mixed Linear Model for continuous data or a logistic Generalised Estimating

Equation for categorical data. Statistical tests were performed using SPSS 15 (SPSS Inc)

and statistical significance was set at p<0.05.

67

2.3 Results

2.3.1 Baseline Details

Baseline patient characteristics are summarised in Table 2.1. Patients with atrial

fibrillation had a median arrhythmia history of 72 months and a mean longest episode

of 2.8±1.9 days. Groups were well matched for age, gender, left ventricular wall

thickness and function. However, patients with atrial fibrillation had significantly larger

left atria (p=0.004). None of the patients with atrial fibrillation had atrial arrhythmias

(>30 seconds) by continuous monitoring in the 7 days prior to the procedure.

2.3.2 Atrial Refractoriness

For patients with atrial fibrillation, the mean ERP across all sites was greater than that

of reference patients (at 600ms: 255±25 versus 222±16ms, p<0.001; at 450ms: 234±20

versus 212±14ms, p=0.004). Figure 2‐2 demonstrates the ERP at the ten sites tested

following 600 and 450ms drive trains in patients with atrial fibrillation versus reference

patients, and indicates those areas where significant differences were seen. The

differences in ERP between patients with atrial fibrillation and reference patients were

not uniform across all sites (p=0.02). At each site, patients with atrial fibrillation

demonstrated either an increase or no significant change in ERP when compared to

the reference group. The ERP was higher following a 600ms drive train than after a

450ms drive train for both patients with atrial fibrillation (p<0.001) and reference

patients (p=0.02); that is, there was preservation of physiological rate‐adaptation of

ERP in both groups (p=0.1).

68

2.3.3 Atrial Conduction Time

The conduction time along the catheters at the left atrial roof, inferior left atrium,

coronary sinus and lateral right atrium was significantly longer in patients with atrial

fibrillation compared to the reference group (57±18 versus 47±10ms, p=0.01).

However, there was no significant interaction between patient group and site of

conduction measurement (p=0.3), suggesting a homogenous slowing of conduction in

patients with atrial fibrillation.

The P wave duration was significantly prolonged in patients with atrial fibrillation

compared to reference patients (128±8 versus 95±10ms, p<0.001).

2.3.4 Site‐Specific Conduction Abnormalities

Site‐specific conduction abnormalities at the crista terminalis during sinus rhythm

were more apparent in patients with atrial fibrillation than reference patients. During

sinus rhythm, patients with atrial fibrillation had a greater number of bipoles

demonstrating double potentials or fractionated signals on the crista terminalis

catheter than reference patients (4.9±1.8 versus 0.7±0.8, p<0.001) of which the

maximum electrogram duration was longer (84±15 versus 52±11ms, p<0.001). These

differences were even more marked during pacing and increased further with extra‐

stimulus (Figures 2‐3 and 2‐4). These data highlight the functional nature of the

conduction delay at the crista terminalis, as evidenced by variation in the extent of

conduction abnormalities depending on the rate and site of stimulation.

69

2.3.5 Sinus Node Function

Patients with atrial fibrillation had a longer baseline sinus cycle length (975±131 versus

762±129ms, p<0.001), SACT (154±58 versus 83±31ms, p<0.001) and CSNRT at 600ms

(265±57 versus 185±60ms, p=0.002) but not 450ms (261±96 versus 241±76ms, p=0.6)

compared to reference patients.

2.4 Discussion

2.4.1 Major findings

This study utilised electrophysiological mapping to demonstrate new information on

the nature of abnormalities within the atria of patients with paroxysmal “lone” atrial

fibrillation, remote from an arrhythmic event.

The major findings are as follows:

(i) Conduction abnormalities characterised by prolongation of conduction

times along linearly placed catheters, longer P wave duration and site‐

specific conduction delay.

(ii) Impaired sinus node function indicated by prolonged SACT and CSNRT.

(iii) An increase in ERP, which is consistent with prior studies evaluating clinical

substrates for atrial fibrillation but is in contrast to the remodelling

attributed to atrial fibrillation itself.

Thus, this suggests that patients with paroxysmal lone atrial fibrillation have an

abnormal atrial electrophysiological substrate. We posit that these abnormalities

70

contribute to the “second factor” that promotes progression of atrial fibrillation and

why sinus rhythm does not beget sinus rhythm.

2.4.2 Rate‐Related Electrical Remodelling

The seminal observations by Wijffels et al. (1995) that “atrial fibrillation begets atrial

fibrillation” has as its central tenet the observation of reduction in atrial ERP by

repeated induction of atrial fibrillation thereby allowing atrial fibrillation to sustain

itself. Many experimental studies have confirmed this finding (Morillo et al. 1995;

Elvan et al. 1996; Goette et al. 1996; Fareh et al. 1998; Lee et al. 1999). While these

initial studies fail to observe any conduction abnormalities, subsequent studies by

Gaspo et al. (1997) demonstrated that progressively longer episodes of atrial

fibrillation also resulted in conduction slowing. Thus, the electrical remodelling of the

atria in response to atrial fibrillation lends toward the perpetuation of arrhythmia by a

shortening of the reentrant “wavelength”. In addition, several investigators have also

observed remodelling of the sinus node that develops soon after the initiation of

arrhythmia and reverses after restoration of sinus rhythm (Kumagai et al. 1991; Elvan

et al. 1996; Hocini et al. 2003).

Clinical studies have also observed similar effects of rate on the atria. Patients with

recent atrial fibrillation have been shown to have shorter ERP than control groups, as

well as sinus node dysfunction and conduction delay (Cosio et al. 1983; Kumagai et al.

1991). Yu et al. (1999) demonstrated shorter ERP with impaired rate adaptation and

depressed conduction following cardioversion for longstanding atrial fibrillation, with

the ERP gradually increasing after 4 days of sinus rhythm. Kojodjojo et al. (2007)

71

studied patients with paroxysmal and persistent atrial fibrillation at their ablation

procedure and demonstrated a reduction in ERP at 3 sites compared to patients with

no atrial fibrillation. In this study, patients with persistent atrial fibrillation had just 10

minutes of sinus rhythm prior to study and those with paroxysmal atrial fibrillation had

a mean atrial fibrillation burden of 3.2±3.7 hours in the preceding 72 hours. However,

with the known experimental evidence of rate‐related remodelling, it is unclear if the

abnormalities observed in such patients with atrial fibrillation form the underlying

substrate for or are a consequence of atrial fibrillation.

2.4.3 Progression of Atrial Fibrillation

Garratt et al. (1999) evaluated the effect of repeated episodes of atrial fibrillation on

atrial electrical remodelling. While marked remodelling was observed with five days of

atrial fibrillation, two days of sinus rhythm was adequate to reverse all abnormalities.

Despite repetitive episodes of atrial fibrillation, there were no additive changes

observed in atrial electrical remodelling. Fynn et al. (2002) evaluated the strategy of

repeated cardioversion of atrial fibrillation in the clinical setting. Early and repeated

cardioversion was performed to maintain sinus rhythm but failed to prevent the

progression to more atrial fibrillation. Indeed, their findings suggest that sinus rhythm

does not beget sinus rhythm in patients with a history of atrial fibrillation.

These findings have led Allessie and co‐workers to propose the existence of a “second

factor” for the initial substrate and subsequent progression of atrial fibrillation (Garratt

et al. 1999). To identify this factor we studied patients with paroxysmal lone atrial

fibrillation who were remote from an episode of atrial fibrillation (thereby removing

72

the effects of acute rate‐related remodelling). When studied distant from an

arrhythmic event, patients with lone atrial fibrillation demonstrate an increase in atrial

refractoriness, widespread and site‐specific conduction abnormalities, and evidence of

sinus node remodelling. We put forward that these changes in conduction seen in lone

paroxysmal atrial fibrillation, when studied remote from arrhythmia, contribute to the

“second factor” which plays an important role in the development and progression of

atrial fibrillation.

2.4.4 Substrate Predisposing to the Development of Atrial Fibrillation

Although the abnormalities that have been observed to result from atrial fibrillation

itself have been proposed as the mechanisms by which atrial fibrillation begets atrial

fibrillation, it is not likely that rate‐related remodelling forms the substrate

predisposing to the development of atrial fibrillation in the first place. This has led

several investigators to evaluate the atrial abnormalities in conditions predisposed to

the development of atrial fibrillation.

Li et al. (1999) developed a canine model of congestive heart failure to evaluate effects

on atrial remodelling and demonstrated "atrial remodelling of a different sort" to that

seen due to atrial fibrillation alone. A significant increase in ERP at short cycle lengths

associated with an increase in the heterogeneity of conduction and marked interstitial

fibrosis was seen. Despite the increase in ERP, these abnormalities resulted in a

significant increase in the duration of atrial fibrillation. Verheule et al. (2003) observed

a propensity for atrial fibrillation despite an increase in ERP in a canine model of mitral

73

regurgitation, again due to abnormalities in conduction and profound structural

remodelling.

Clinical studies of the atrial substrate in patients known to be predisposed to atrial

fibrillation but without antecedent arrhythmia have demonstrated similar

observations. Morton et al. (2003) studied the right atrial substrate in patients with

atrial septal defects and observed prolonged ERP and delayed conduction across the

crista terminalis compared to reference patients. Sanders et al. showed similar findings

of prolonged ERP with widespread and site‐specific conduction abnormalities in

patients with congestive heart failure (2003) and sinus node disease (2004a). Kistler et

al. (2004) demonstrated such changes in a progressive manner with increasing age.

While these findings have been restricted to right atria, comparable results have

recently been observed in the left atria of patients with mitral stenosis (John et al.

online early).

In patients with atrial fibrillation there is a paucity of clinical electrophysiological

studies performed outside the context of recent arrhythmia. In order to minimise the

confounding effects of electrical remodelling, we have undertaken meticulous

screening to identify patients who were in sinus rhythm for a minimum of 7 days prior

to their ablation procedure. Using such stringent enrolment criteria, the observations

of the current study are remarkably consistent with the findings observed in the above

conditions on the substrate predisposing to atrial fibrillation. Taken together, these

studies provide compelling evidence that the significant contributors giving rise to the

substrate underlying atrial fibrillation are not the changes in refractoriness, but rather

the conduction abnormalities discussed above.

74

2.4.5 Implications

The electrical remodelling seen in conjunction with ongoing arrhythmia and thought to

perpetuate atrial fibrillation is known to normalise within days of sinus rhythm (Garratt

et al. 1999). However, strategies of repeated cardioversion have not been successful in

retarding progression of disease (Fynn et al. 2002). By studying patients remote in time

from the arrhythmia itself, we have been able to show that a significant contributor to

the underlying substrate in patients with atrial fibrillation is in fact that of conduction

abnormalities, which may account for the progression of the atrial fibrillation disease.

Future strategies to treat atrial fibrillation should therefore focus on modification of

atrial conduction. Finally, while pulmonary vein isolation is currently considered to be

highly successful for the majority of patients with paroxysmal atrial fibrillation, there is

yet to be definitive long‐term data on outcome. This study showing an abnormal

substrate in lone atrial fibrillation raises the possibility of progressive disease

continuing despite early procedural success.

2.4.6 Limitations

While the abnormalities observed in this study are proposed to constitute the

substrate predisposing to atrial fibrillation, the development of clinical atrial fibrillation

is complex and depends not only on substrate but also on other factors such as triggers

and perpetuators that were not addressed by this study. Patients with persistent atrial

fibrillation were not studied and may differ in substrate, however many patients with

persistent atrial fibrillation start with paroxysms initially. Finally, while we monitored

patients for a week prior to the procedure to ensure our evaluation was remote from

75

an episode of atrial fibrillation, we cannot exclude an effect of rate‐related remodelling

from previous episodes. However, prior studies have demonstrated no additive effects

of episodic atrial fibrillation and resolution of remodelling within days (Garratt et al.

1999).

2.5 Conclusion

Patients with paroxysmal lone atrial fibrillation, remote from an arrhythmic episode,

demonstrate electrophysiological abnormalities characterised by widespread and site‐

specific conduction abnormalities, altered sinus node function and prolonged atrial

refractoriness. These abnormalities contribute to the “second factor” critical to the

development and progression of atrial fibrillation.

76

Table 2.1

Patient characteristics

AF (n=13)

Reference (n=13)

p value

Age (years) 54±11 53±8 0.7

Male gender (%) 69 46 0.4

Median AF history (months) 72 0 n/a

Longest AF duration (days) 2.8±1.9 0 n/a

Left atrial parasternal size (mm) 42±7 34±5 0.004

Interventricular septum (mm) 11±2 11±1 0.7

Left Ventricular ejection fraction (%) 54±9 58±8 0.7

77

Figure 2‐1

Catheter positions on fluoroscopy

Left anterior oblique views of a decapolar catheter in the coronary sinus, a

duodecapolar catheter along the crista terminalis with its stabilising sheath, a

duodecapolar catheter along the lateral right atrium and a decapolar catheter with its

stabilising sheath in the left atrium. The left atrial catheter was sequentially positioned

by looping the catheter to record along the left atrial roof (left) and then withdrawn to

record the inferior left atrium (right). Subsequently, the decapolar catheter was

positioned at the left atrial appendage, posterior left atrial wall and the high‐septal

right atrium (not shown).

78

Figure 2‐2

Regional effective refractoriness

Figure legend overleaf.

79

Figure 2‐2 (previous page)

Mean (+ standard deviation) ERP at the ten sites tested following a 600ms drive train

(top) and a 450ms drive train (bottom) in patients with atrial fibrillation and reference

patients. All values for 600ms are higher than the same patient group and site value

for 450ms. The difference in mean ERP between patients with atrial fibrillation and

reference patients varied significantly according to the region of measurement

(p=0.02).

*p<0.001, †p=0.01, ‡p<0.05 for post‐hoc comparisons.

dCS – distal coronary sinus

pCS – proximal coronary sinus

LLRA – low lateral right atrium

HLRA – high lateral right atrium

HRAS – high right atrial septum

LSPV – junction of left superior pulmonary vein with left atrium

RSPV – junction of right superior pulmonary vein with left atrium

InfLA – inferior left atrium. LAA – left atrial appendage

PostLA – posterior left atrium.

80

Figure 2‐3

Conduction characteristics at the crista terminalis

Number of abnormal electrograms

Mean (+ standard deviation) number of double potentials or fractionated signals along

the crista terminalis during pacing at 600ms (S1, top) and with the earliest captured

extra‐stimulus (S2, bottom) from 4 sites in the atria in patients with atrial fibrillation

and reference patients. *p<0.001, †p<0.01. pCS – proximal coronary sinus. LLRA – low

lateral right atrium. InfLA – inferior left atrium. LAA – left atrial appendage.

81

Figure 2‐4


Maximum duration of abnormal electrograms

Mean (+ standard deviation) of the maximum duration of double potentials or

fractionated signals along the crista terminalis during pacing at 600ms (S1) and with

the earliest captured extra‐stimulus (S2) from 4 sites in the atria in patients with atrial

fibrillation and reference patients. *p<0.001, †p<0.01. pCS – proximal coronary sinus.

LLRA – low lateral right atrium. InfLA – inferior left atrium. LAA – left atrial appendage.

82

Chapter 3. Right and left atrial electroanatomical substrate of lone atrial

fibrillation‐‐‐‐

3.1 Introduction

The initiation and maintenance of atrial fibrillation is brought about by intricate

interactions between triggers and perpetuators within the atrial substrate (Allessie et

al. 2001). The perpetuation of atrial fibrillation is aided by the electrical remodelling

undergone by the atria and demonstrated repeatedly in the experimental setting

(Morillo et al. 1995; Wijffels et al. 1995; Goette et al. 1996; Gaspo et al. 1997; Fareh et

al. 1998). However, while this electrical remodelling is seen clinically, the effects are

reversible once sinus rhythm is restored (Yu et al. 1999; Hobbs et al. 2000).

Nevertheless, policies of early and repeated cardioversion to avoid self‐sustaining

arrhythmia have not halted the progression of disease (Fynn et al. 2002). This

observation has prompted the hunt for a “second factor” promoting the evolution and

advancement of atrial fibrillation.

The structural and electrical changes associated with atrial fibrillation are difficult to

differentiate from that due to the affiliated conditions that may be giving rise to the

arrhythmia; conditions such as heart failure and hypertension. Patients with “lone”

atrial fibrillation however, are defined as a group without overt concomitant

conditions (Kopecky et al. 1987; Kumagai et al. 1991; Jaïs et al. 2000) and study of

these individuals may provide insight into the underlying substrate of the arrhythmia

itself. Furthermore, the arrhythmia alone can lead to alterations in the atrial structure

which may tend to perpetuate atrial fibrillation (Ausma et al. 2003). Studying patients

with lone atrial fibrillation, after allowing sufficient time for electrical remodelling to

83

reverse, would allow conclusions to be drawn about the structural and electrical

changes seen in the atrial substrate predisposing to atrial fibrillation, without the

effects of associated heart disease and rate‐related remodelling.

We hypothesised that the disease progression of patients with paroxysmal “lone”

atrial fibrillation is determined by abnormal atrial substrate. In order to evaluate the

substrate predisposing to atrial fibrillation, we evaluated the electroanatomical

characteristics of atria in patients with paroxysmal “lone” atrial fibrillation, remote in

time from the arrhythmia.

3.2 Methods


This study comprised 12 patients undergoing first‐time ablation for paroxysmal atrial

fibrillation and a reference group of 12 patients with structurally normal hearts

undergoing radiofrequency ablation for atrioventricular reentry tachycardia with left‐

sided accessory pathways and no history of atrial fibrillation. Patients in the study

group were excluded if they had atrial fibrillation during the week prior to ablation

(established by continuous monitoring) or any of the criteria that would prohibit the

diagnosis of “lone” atrial fibrillation, previously defined as the absence of structural


blood chemistry, and trans‐thoracic as well as trans‐oesophageal echocardiography

(Kopecky et al. 1987; Kumagai et al. 1991; Jaïs et al. 2000). Coronary artery disease was

excluded by clinical, ECG or stress test criteria. Pulmonary disease, hypertension,

84

hyperthyroidism and diabetes were eliminated by appropriate tests. Paroxysmal atrial

fibrillation was defined according to the Expert Consensus Statement as recurrent

atrial fibrillation that terminates spontaneously within 7 days (Calkins et al. 2007).

Thus, of 106 patients with highly symptomatic atrial fibrillation presenting for ablation

screened, 12 patients were suitable to be included in this study.




Adelaide Hospital.



utilising midazolam and fentanyl. In patients with atrial fibrillation the study protocol

was performed before ablation, while in the reference group this was done after

accessory pathway ablation. The left atrium was accessed using a single transeptal

puncture following which repeated bolus unfractionated heparin was utilised to

maintain the activated clotting time between 300‐350 seconds.

Following the study protocol, patients with atrial fibrillation underwent circumferential

pulmonary vein ablation with an end‐point of isolation confirmed by circumferential

mapping (Lasso, Biosense‐Webster) with either elimination or dissociation of

pulmonary venous potentials. Ablation of the pulmonary veins was performed using a

delivered power of 30W with irrigation rates of 30mL/min (Thermocool, Biosense

Webster). Additional substrate modification was performed in patients with an

85

episode of atrial fibrillation ≥48 hours or with a left atrial size ≥57mm (longest

diameter). This took the form of linear ablation along the left atrial roof with an end‐

point of linear double potentials and conduction block demonstrated by an activation

detour during pacing of the left atrial appendage. Cavotricuspid isthmus ablation with

an endpoint of bi‐directional isthmus block was performed only in patients with a

history of typical flutter or if mapping confirmed typical flutter during the procedure.

Linear ablation was performed with a delivered power of 30‐35W with irrigation rates

of 30‐60mL/min.

3.2.3 Electroanatomical Study Protocol

Electroanatomical maps were created of both atria during sinus rhythm using the

CARTO mapping system (Biosense‐Webster). The electroanatomical mapping system

has been previously described in detail; the accuracy of the sensor position has been

previously validated and is 0.8mm and 5 degrees (Gepstein et al. 1997). In brief, the

system records the surface ECG and bipolar electrograms filtered at 30‐400Hz from the

mapping and reference catheters. Endocardial contact during point acquisition was

facilitated by electrogram stability, fluoroscopy and the catheter icon on the CARTO

system. Points were acquired in the auto‐freeze mode if the stability criteria in space

(≤6mm) and local activation time (≤5ms) were met. Mapping was performed with an

equal distribution of points using a fill‐threshold of 15mm. Editing of points was

performed off‐line. Local activation time was manually annotated to the peak of the

largest amplitude deflection on bipolar electrograms; in the presence of double

potentials this was annotated at the largest potential. If the bipolar electrogram

86

displayed equivalent maximum positive and negative deflections, the maximum

negative deflection on the simultaneously acquired unipolar electrogram was used to

annotate the local activation time. Points not conforming to the surface ECG P wave

morphology or <75% of the maximum voltage of the preceding electrogram were

excluded. Regional atrial bipolar voltage and conduction velocity were analysed as

previously described and are detailed below (Sanders et al. 2003; Kistler et al. 2004;

Sanders et al. 2004a; John et al. online early).

3.2.3.1 Voltage Analysis

For the purposes of evaluating regional voltage differences, each atrium was

segmented using previously validated offline software (Kuklik et al. 2004). The right

atrium was segmented as the high‐ and low‐lateral right atrium, high‐ and low‐

posterior right atrium, high‐ and low‐septal right atrium, and anterior right atrium. The

left atrium was segmented as posterior left atrium, anterior left atrium, septal left

atrium, inferior left atrium, lateral left atrium and left atrial roof. The voltage of points

identified by region was then exported for analysis. Low voltage points were defined as

point with a bipolar voltage ≤0.5mV and electrically silent points as the absence of

recordable activity or a bipolar voltage amplitude ≤0.05mV (the noise level of the

system).

3.2.3.2 Conduction Velocity Analysis

Isochronal activation maps (5ms intervals) of the atria were created and regional

conduction velocity determined in the direction of the wave‐front propagation (least

isochronal crowding). An approximation of conduction velocity was determined by

expressing the distance between two points as a function of the difference in local

87

activation time. Mean conduction velocity for each region was determined by

averaging the conduction velocity between 5 pairs of points, as previously described

(Sanders et al. 2003; Kistler et al. 2004; Sanders et al. 2004a; John et al. online early).

For the purposes of evaluating regional conduction differences, each atrium was

segmented as above.

3.2.3.3 Complex Electrograms

The proportion of points demonstrating complex electrograms was determined using

the following definitions: (i) Fractionated signals – complex activity of ≥50ms duration;

and (ii) Double potentials – potentials separated by an isoelectric interval where total

electrogram duration was ≥50ms.





between means were analysed using paired or unpaired t tests as appropriate.

Comparisons with adjustment for multiple sampling within patients were performed

using a Mixed Linear Model for continuous data or a logistic Generalised Estimating

Equation for categorical data. Statistical tests were performed using SPSS 15 (SPSS Inc)

and statistical significance was set at p<0.05.

88

3.3 Results


Baseline patient characteristics are summarised in Table 3.1. Patients with atrial

fibrillation had a median arrhythmia history of 60 months and a mean longest episode

of 3.2±2.6 days. Groups were well matched for age, gender, left ventricular wall

thickness and function. However, patients with atrial fibrillation had significantly larger

left atria (p=0.01). None of the patients with atrial fibrillation had atrial arrhythmias

(>30 seconds) by continuous monitoring in the 7 days prior to the procedure.

A total of 193±62 points per patient were analysed in the left atrium and right atrium

using electroanatomical mapping.

3.3.2 Structural and Voltage Abnormalities

Both right and left atrial volumes by electroanatomical mapping were significantly

greater in patients with atrial fibrillation compared to reference patients; right atrium:

94±18 versus 69±9mL, p=0.003 and left atrium: 99±19 versus 77±17mL, p=0.006.

The mean bipolar voltage was reduced in patients with atrial fibrillation compared to

reference patients (right atrium: 1.7±0.4 versus 2.9±0.4mV; left atrium: 1.7±0.7 versus

3.3±0.7mV, p<0.001). The difference in mean log bipolar voltage between patients

with atrial fibrillation and reference patients varied significantly by Mixed Linear model

according to the region of measurement (right atrium: p=0.002; left atrium: p=0.03).

Regional differences in bipolar voltage with significant differences indicated are

illustrated in Figure 3‐1. Additionally, points in patients with atrial fibrillation at the

89

high‐lateral right atrium, posterior left atrium and the left atrial roof were more likely

to be low voltage (≤0.5mV) (OR 2.9, 95% CI 1.4‐6.3; OR 1.7, 95% CI 1.1‐2.6; OR 3.3, 95%

CI 1.8‐6.3, respectively). No significant areas of electrical silence (≤0.05mV) were

observed in this cohort of patients. Representative examples of electroanatomical

maps in representative patients are shown in Figure 3‐2.

3.3.3 Abnormalities in Conduction Velocity

Patients with atrial fibrillation had a significantly slower mean conduction velocity

during sinus rhythm compared to reference patients (right atrium: 1.3±0.3 versus

2.1±0.5mm/ms; left atrium: 1.2±0.2 versus 2.2±0.4mm/ms, p<0.001). Regional

differences in conduction velocity with significant differences indicated are illustrated

in Figure 3‐3. The total atrial activation time was significantly prolonged in patients

with atrial fibrillation compared to reference patients (128±17 versus 89±10ms,

p<0.001).

3.3.4 Complex Electrograms

Patients with atrial fibrillation demonstrated a significantly greater number of points

with double potentials or fractionated signals than reference patients (27±8 versus

8±5%, p<0.001). Patients with atrial fibrillation were more likely to have a point with a

double potential or fractionated signal in any region than reference patients (p<0.02).

These points were distributed throughout both atria with a clustering at the high‐

90

posterior and high‐septal regions of the right atrium, and the left atrial septum and

roof (Figure 3‐2).

3.4 Discussion


This study utilised electroanatomical mapping to demonstrate new information on the

nature of abnormalities within the atria of patients with paroxysmal “lone” atrial

fibrillation, remote from an arrhythmic event.


(i) Structural abnormalities characterised by atrial dilatation and lower mean

atrial voltage suggesting the loss of atrial myocardium.

(ii) Slower conduction velocities.

(iii) An increase in the proportion of electrograms characterised as double

potentials or fractionated signals.

Thus, these findings suggest that patients with paroxysmal lone atrial fibrillation have

an abnormal electroanatomical atrial substrate. We posit that these abnormalities,

contribute to the “second factor” that promotes progression of atrial fibrillation and

why sinus rhythm does not beget sinus rhythm.

91

3.4.2 Structural changes in Atrial Fibrillation

Despite the early occurrence and reversibility of electrical remodelling of the atria in

response to rate, Ausma et al. (2003) have observed that ultrastructural changes

evolve over a much longer time frame and are not completely resolved as late as four

months after return to sinus rhythm from prolonged atrial fibrillation. These slowly

resolving alterations have been characterised as an increase in myolysis and the

fraction of extracellular matrix per myocyte, as well as changes in the expression of

structural proteins. These observations were associated with longer episodes of

induced atrial fibrillation when compared to baseline inductions.

Recent studies in patients with atrial fibrillation undergoing ablation have

demonstrated the presence of areas of low voltage and electrical silence (Lo et al.

2007; Marcus et al. 2007), and have implicated this substrate as a marker of a worse

outcome following an ablation procedure (Verma et al. 2005). However, these studies

have included patients with structural heart disease and recent arrhythmia, and

therefore may reflect the effects of the associated heart disease and/or remodelling

due to atrial fibrillation itself.

3.4.3 Electroanatomical Studies in Atrial Fibrillation and Related Conditions

In patients with atrial fibrillation there has been little clinical mapping performed

outside the context of recent arrhythmia. For example, Kojodjojo et al. (2007) studied

the electroanatomical maps of patients with paroxysmal and persistent atrial

fibrillation created at their ablation procedure and demonstrated slowing of wave‐

front propagation velocity compared to patients with no atrial fibrillation. In this study,

92

patients with persistent atrial fibrillation had just 10 minutes of sinus rhythm prior to

study and those with paroxysmal atrial fibrillation had a mean arrhythmia burden of

3.2±3.7 hours in the preceding 72 hours. It is therefore uncertain how much of this

observed change is due to electrical remodelling from recent arrhythmia and how

much it is reflecting the actual underlying substrate.

One of the mechanisms by which atrial fibrillation begets atrial fibrillation has been

suggested to be rate‐related remodelling. However, it is unlikely that the abnormalities

that have been observed to result from atrial fibrillation itself form the substrate

predisposing to the development of atrial fibrillation in the first instance. Insights into

the underlying substrate can be gained from looking at the atrial abnormalities

reported in conditions known to be predisposed to atrial fibrillation. Many of these

studies have used electroanatomical maps to illustrate such characteristics in patients

with conditions associated with atrial fibrillation, but without atrial fibrillation

themselves. Sanders et al. studied the right atrial substrate of patients with congestive

heart failure (2003) and sinus node disease (2004a) and reported areas of low voltage

and electrical silence in association with conduction abnormalities. Kistler et al. (2004)

demonstrated progressive changes in older patients without significant heart disease,

when compared to younger cohorts of patients. While these findings have been

restricted to the right atrium, recently similar characteristics have been observed via

electroanatomical mapping in the left atria of patients with mitral stenosis (John et al.

online early).

93

3.4.4 Prior Studies Implicating Abnormalities in “Lone” Atrial Fibrillation

Although patients with “lone” atrial fibrillation are defined as a group with no

structural heart disease, there is accumulating evidence of occult abnormalities.

Frustaci et al. (1997) found abnormal atrial histology including patchy fibrosis and

inflammatory infiltrates in all of 12 patients with paroxysmal lone atrial fibrillation, and

in none of 11 control patients. The fibrosis has been characterised as an increased

concentration of collagen type I in patients with lone atrial fibrillation compared to

controls and this increase has been observed to be attenuated in those patients

receiving angiotensin converting enzyme inhibitors (Boldt et al. 2006). There is further

evidence that inflammation plays a role in atrial fibrillation, with the observation that

C‐reactive protein is elevated in patients with lone atrial fibrillation (Chung et al. 2001).

Jaïs et al. (2000) have demonstrated by transeptal haemodynamic evaluation that

diastolic dysfunction is present in patients with lone atrial fibrillation. Finally, there is

some evidence to suggest that patients with lone atrial fibrillation have microvascular

dysfunction in the coronary circulation (Skalidis et al. 2008). Thus, it is feasible that

patients without overt structural heart disease may have fibrosis, inflammation,

structural change and vascular dysfunction; all features that are consistent with and

contribute to our observations of an abnormal substrate in “lone” atrial fibrillation.

3.4.5 Implications

Our understanding of the atrial substrate predisposing to atrial fibrillation is extended

by this study. We have been able to show that the predominant underlying substrate

in patients with atrial fibrillation is slowing in regional conduction, together with

94

structural change reflected by lower voltage and larger atria. Importantly, this data is

from patients without overt structural heart disease and remote in time from the

remodelling effect of recent arrhythmia in order to minimise the effects of associated

heart conditions and the arrhythmia itself. Future strategies to treat atrial fibrillation

should therefore focus on atrial myocardial structure and conduction. While there is

yet to be definitive long‐term data on the outcome of ablation treatments for patients

with paroxysmal atrial fibrillation, this study raises the possibility of progressive

disease continuing despite apparent procedural success due to the continuing

evolution of abnormal substrate in lone atrial fibrillation.

3.4.6 Limitations

Patients with persistent atrial fibrillation were not studied and may differ in substrate,

however many patients with persistent atrial fibrillation have progressed from

paroxysms initially. While the abnormalities observed in this study are proposed to

constitute the substrate predisposing to atrial fibrillation, the development of clinical

atrial fibrillation is complex and depends not only on substrate but also on other

factors such as triggers and perpetuators that were not addressed by this study.

3.5 Conclusion

Patients with paroxysmal lone atrial fibrillation demonstrate structural abnormalities

characterised by loss of myocardial voltage, conduction slowing and an increase in the

95

proportion of complex electrograms. These abnormalities contribute to the “second

factor” critical to the development and progression of atrial fibrillation.

96

Table 3.1


AF (n=12)

Reference (n=12)

p value

Age (years) 51±7 47±9 0.2

Male gender (%) 92 75 0.6

Median AF history (months) 60 0 n/a

Longest AF duration (days) 3.0±2.6 0 n/a



Left Ventricular ejection fraction (%) 55±9 57±9 0.8

97

Figure 3‐1

Regional bipolar voltage

Mean (+ standard deviation) bipolar voltage of the 7 right and 6 left atrial regions from

the electroanatomical map.

*p<0.001, †p<0.01, ‡p<0.05. RA – right atrium. LA – left atrium. H – high. L – low.

98

Figure 3‐2

Example electroanatomical bipolar voltage maps

Representative CARTO maps of a patient with atrial fibrillation (bottom) and a

reference patient (top). Both atria are oriented in the posterior‐anterior projection and

are of the same scale. The colour scale is identical in both images with red

representing voltage ≤0.5mV and purple being voltage ≥5mV. In addition to having

greater regions of low voltage (red), the patient with atrial fibrillation has more

evidence of conduction abnormalities in the form of fractionated signals (pink tags)

and double potentials (blue tags).

99

Figure 3‐3

Regional conduction velocity

Mean (+ standard deviation) conduction velocity in the 7 right and 6 left atrial regions

from the electroanatomical map.

*p<0.001, †p=0.01, ‡p<0.05. RA – right atrium. LA – left atrium. H – high. L – low.

100

Chapter 4. Right atrial electrical substrate of atrial flutter

4.1 Introduction

The circuit of typical atrial flutter has been well characterised with the crista terminalis

and the tricuspid annulus defined as the posterior and anterior conduction barriers,

respectively (Cosio et al. 1988; Olgin et al. 1995; Kalman et al. 1996; Nakagawa et al.

1996). Ablation of the cavotricuspid isthmus is a highly successful treatment for atrial

flutter, however a significant proportion subsequently develop atrial fibrillation

(Paydak et al. 1998; Chinitz et al. 2007; Ellis et al. 2007). Indeed, atrial flutter has a

close inter‐relationship with atrial fibrillation and the two rhythms frequently co‐exist

(Halligan et al. 2004). Key to the understanding of this relationship are the seminal

observations by Waldo and colleagues that typical atrial flutter almost always develops

from antecedent atrial fibrillation of variable duration (Shimizu et al. 1991; Waldo and

Cooper 1996). One of the postulates for this observation has been that a period of

atrial fibrillation facilitates the development of the functional conduction block along

the crista terminalis required for atrial flutter to sustain.

It is well recognised that atrial fibrillation occurs in patients with structural heart

disease and an abnormal atrial substrate (Kannel et al. 1982b). The atrial substrate has

been extensively characterised in conditions predisposed to the development of atrial

fibrillation such as heart failure, hypertension, valvular heart disease, congenital heart

disease, sinus node disease and ageing (Li et al. 1999; Morton et al. 2003; Sanders et

al. 2003; Verheule et al. 2003; Kistler et al. 2004; Sanders et al. 2004a; Kistler et al.

2006; John et al. online early). Whether similar such abnormal atrial substrate exists in

101

patients with atrial flutter to account for the frequent co‐existence of both

arrhythmias is not known.

We hypothesised that patients with atrial flutter would demonstrate an abnormal

atrial substrate to account for the frequent development of atrial fibrillation. In order

to evaluate the substrate without the electrical remodelling effects of the arrhythmia

itself, we performed electrophysiological mapping of right atrium remote in time from

the arrhythmia.

4.2 Methods


This study comprised 8 patients without antecedent atrial fibrillation undergoing

ablation for typical atrial flutter and a reference group of 12 patients undergoing

radiofrequency ablation for atrioventricular reentry tachycardia or atrioventricular

nodal reentry tachycardia. Patients with atrial flutter were selected on the basis of

having ≥2 previous episodes of typical atrial flutter confirmed on a 12‐lead

electrocardiogram. These patients were documented to be in sinus rhythm and

described no symptoms attributable to atrial flutter for at least 1 month prior to

ablation. In addition, all patients underwent continuous monitoring for 7 days prior to

the study and were excluded if they had evidence of any atrial arrhythmia >30

seconds.



102


Adelaide Hospital.



utilising midazolam and fentanyl. In patients with atrial flutter the study protocol was

performed before ablation, while in the reference group this was done after ablation.

Following the study protocol, patients with atrial flutter underwent cavotricuspid

isthmus ablation with a delivered power of 35W with irrigation rates of 30‐60mL/min

and an endpoint of bi‐directional isthmus block.

4.2.3 Electrophysiology Study Protocol

The following catheters were positioned for the study protocol: (i) a 10‐pole catheter

(2‐5‐2mm inter‐electrode spacing, Daig Electrophysiology) within the coronary sinus

with the proximal bipole at the coronary sinus ostium as determined in the best septal

left anterior oblique position; (ii) a 20‐pole “crista” catheter (1‐3‐1mm inter‐electrode

spacing, Biosense‐Webster) placed along the crista terminalis with the distal tip

superiorly such that the second bipole lay at the junction of the superior vena cava and

right atrium, stabilised by a long sheath (CSTA, Daig Electrophysiology) to ensure close

apposition to the postero‐lateral right atrium; and (iii) a 20‐pole catheter (2‐5‐2mm

inter‐electrode spacing, Daig Electrophysiology) placed initially along the lateral right

103

atrium and then moved to the high right atrial septum (Figure 4‐1). The following

parameters were determined:

4.2.3.1 Effective Refractoriness

Atrial ERP was evaluated at twice diastolic threshold at cycle lengths of 600 and 450ms

using an 8‐beat drive followed by an extra‐stimulus, starting with a coupling interval of

150ms increasing in 10ms increments. ERP was defined as the longest coupling interval

failing to propagate to the atrium. At each site the ERP was measured 3 times at each

cycle length and averaged. If ERP varied by >10ms an additional 2 measurements were

made and the total number averaged. ERP was measured from the following sites: (i)

distal coronary sinus; (ii) proximal coronary sinus; (iii) low‐lateral right atrium; (iv) high‐

lateral right atrium; and (v) high‐septal right atrium. Atrial fibrillation induced by ERP

testing lasting >5 minutes was considered sustained; when this occurred, no further

data was acquired.

4.2.3.2 Atrial Conduction

Atrial conduction time was assessed along linearly placed catheters by pacing

the distal bipole (1,2) and determining the conduction time to a proximal bipole (9,10)

at the coronary sinus and lateral right atrium. Conduction time at each site was

averaged over 10 beats during stable capture at 600 and 450ms cycle lengths.

P wave duration was determined as a marker of global conduction by the

average of 10 beats on ECG lead II.

104

4.2.3.3 Site‐specific Conduction Abnormalities at the Crista Terminalis

The number of bipoles on the crista terminalis catheter with discrete double potentials

separated by an isoelectric interval or complex fractionated activity of ≥50ms duration,

and the maximum electrogram duration were determined during sinus rhythm,

constant pacing and for the shortest‐coupled captured extra‐stimulus from the

proximal coronary sinus and low‐lateral right atrium.

4.2.3.4 Sinus Node Function

Sinus node function was evaluated as follows: (i) Baseline sinus cycle length was

determined over 10 consecutive sinus cycles; (ii) SACT was determined after an 8‐beat

pacing train using the formula [SACT = (return ‐ basic cycle length)/2] 3 times and

averaged; and (iii) CSNRT was determined after a 30‐second drive train at cycle lengths

of 600 and 450ms, correcting for the baseline cycle length. At each cycle length, CSNRT

was determined 3 times and averaged.





between means were analysed using analysis of variance and paired or unpaired t tests

as appropriate. Comparisons with adjustment for multiple sampling within patients

were performed using a Mixed Linear Model for continuous data or a logistic

105

Generalised Estimating Equation for categorical data. Statistical tests were performed

using SPSS 15 (SPSS Inc) and statistical significance was set at p<0.05.

4.3 Results


Baseline patient characteristics are summarised in Table 4.1. Patients with atrial flutter

had a median atrial flutter history of 6.5 months (IQR 3‐13), however they had no

history of any arrhythmia for at least 1 month (range 1‐8 months) prior to the study.

Groups were well matched for age, left ventricular dimensions and hypertrophy, while

differences in atrial size did not reach significance on trans‐thoracic echocardiogram.

4.3.2 Atrial Refractoriness

For patients with atrial flutter, the mean ERP across all sites trended to be greater than

that of reference patients (at 600ms: 244±25 versus 227±15ms, p=0.07; at 450ms:

232±29 versus 213±12ms, p=0.06). At each site patients with atrial flutter

demonstrated equal or higher ERP when compared to the reference group, however

this difference was only statistically significant at one site, the high‐lateral right atrium

(Figure 4‐2). The ERP was higher following 600ms drive trains than after 450ms drive

trains for both patients with atrial flutter (244±25 versus 235±30, p=0.02) and

reference patients (227±15 versus 213±12, p<0.001); no difference in physiological

rate‐adaptation of ERP was observed between groups (p=0.7).

106

4.3.3 Atrial Conduction Time

The conduction time along linear catheters at the lateral right atrium and the coronary

sinus was significantly longer in patients with atrial flutter when compared with

reference patients (p=0.04) and this difference depended on the site of measurement

(p<0.001 for interaction). Post hoc tests performed within the statistical model

revealed slower conduction observed along the lateral right atrium in patients with

atrial flutter than the reference group (67±4 versus 47±3ms, p<0.001) and no

difference in conduction time at the coronary sinus (45±4 versus 46±3ms, p=0.8).

The P wave duration was significantly prolonged in patients with atrial flutter

compared to reference patients (122±18 versus 102±11ms, p=0.007).

4.3.4 Site‐Specific Conduction Abnormalities

Site‐specific conduction abnormalities at the crista terminalis during sinus rhythm

were more apparent in patients with atrial flutter than in reference patients. During

sinus rhythm, patients with atrial flutter had a greater number of bipoles

demonstrating double potentials or fractionated signals on the crista terminalis

catheter than reference patients (4.1±2.6 versus 1.0±1.1 respectively, p=0.001) of

which the maximum electrogram duration was longer (80±26 versus 53±7ms

respectively, p=0.004). These differences were even more marked during pacing and

increased further with extra‐stimulus (Figure 4‐3). These data highlight the functional

nature of the conduction delay at the crista terminalis, as evidenced by variation in the

extent of conduction abnormalities depending on the rate and site of stimulation.

107

4.3.5 Sinus Node Function

Patients with atrial flutter had a longer baseline sinus cycle length than reference

patients (1019±184 versus 777±137ms, p=0.003) and longer CSNRT at 450ms (318±71

versus 203±94ms, p=0.02). Differences in SACT (107±53 versus 82±35ms, p=0.2) and

CSNRT at 600ms (266±113 versus 194±74ms, p=0.1) did not reach significance.

4.4 Discussion

4.4.1 Major Findings

This study utilised electrophysiological mapping to demonstrate new information on

the nature of atrial abnormalities in patients with typical atrial flutter, remote from an

arrhythmic event. The major findings are as follows:

(i) Conduction abnormalities characterised by prolongation of lateral right

atrial conduction time, longer P wave duration and site‐specific conduction

delay.

(ii) Impairment of sinus node function.

(iii) An increase or no change in ERP, consistent with prior studies evaluating

clinical substrates for atrial arrhythmia but in contrast to the remodelling

attributed to atrial arrhythmia itself.

Thus, the present study suggests that patients with atrial flutter, studied remote from

an arrhythmic event, have widespread and persistent abnormal atrial substrate. We

108

posit that these diffuse atrial abnormalities are partly responsible for the frequent

development of atrial fibrillation in patients with atrial flutter.

4.4.2 The Inter‐relationship between Atrial Flutter and Atrial Fibrillation

Ablation of atrial flutter has been observed to significantly reduce the occurrence of

atrial fibrillation at 496±335 days post‐ablation from 55% to 33% (Schmieder et al.

2003). Despite this reduction, reports from patient cohorts followed up after ablation

for atrial flutter describe subsequent atrial fibrillation rates ranging from 25% after

20±9 months surveillance (Paydak et al. 1998) to 82% at 39±11 months follow up (Ellis

et al. 2007). Thus, it is clear that although atrial flutter may be the only clinical

arrhythmia initially apparent, progression to atrial fibrillation is not uncommon.

The atrial arrhythmias of atrial flutter and atrial fibrillation share many of the same

aetiological factors and often co‐exist in the same patient (Vidaillet et al. 2002). Liu et

al. (2002) observed that ablation of conduction gaps in the crista terminalis can result

in the organisation of atrial fibrillation to atrial flutter. Shimizu et al. (1991)

demonstrated in dogs that the onset of atrial flutter in the sterile pericarditis model is

preceded by atrial fibrillation. Ortiz et al. (1994) provided insights into the mechanism

of atrial flutter by demonstrating that maintenance of stable atrial flutter was directly

dependent on sufficient length of functional block being present in the right atrial free

wall. The further observation that triggers in post‐bypass patients led to transitional

atrial fibrillation and then atrial flutter supported the notion that atrial fibrillation,

however short‐lived, was integral to the development of the necessary functional

block (Waldo and Cooper 1996). Roithinger et al. (1997) subsequently observed regular

109

right atrial activation during atrial fibrillation which slowed and organised into atrial

flutter, thereby implicating atrial fibrillation in the generation of atrial flutter.

Testing the hypothesis that atrial fibrillation was indeed necessary for subsequent

atrial flutter, Wazni et al. (2003) studied the effect of trigger elimination by performing

pulmonary vein isolation without cavotricuspid isthmus ablation in 59 patients with

both atrial fibrillation and atrial flutter and found that although 32 of these patients

developed episodes of atrial flutter acutely, only 3 had sustained atrial flutter after 8

weeks. Indeed, patients undergoing ablation for atrial fibrillation in whom atrial flutter

has not been documented clinically or at electrophysiology study have no difference in

outcome with or without cavotricuspid isthmus line (Shah et al. 2007). Patients with

typical atrial flutter in addition to paroxysmal atrial fibrillation have been shown to

have more frequent recurrence of arrhythmia following pulmonary vein cryoisolation

with accompanying isthmus line than patients with atrial fibrillation alone, suggesting

that the substrate of patients with atrial flutter harbours important non‐pulmonary

vein sites of arrhythmogenesis (Moreira et al. 2007).

4.4.3 Substrate Predisposing to the Development of Atrial Flutter

Experimental models have shown the development of an abnormal electrical substrate

with atrial flutter. Morton et al. (2002) demonstrated that prolonged atrial flutter

resulted in an abrupt reduction in refractoriness and a slower reduction in conduction

velocity. Additionally, this study underlined the importance of critically‐timed triggers

in the conversion of atrial flutter to atrial fibrillation after the necessary substrate was

allowed to develop after 4 weeks of atrial flutter.

110

Patients with chronic atrial flutter have been observed to have electrical remodelling

by way of shortened monophasic action potential durations with impaired rate

adaptation compared to control patients (Franz et al. 1997). Ramanna et al. (2005)

have shown that an increase in the dispersion of refractoriness predicts the inducibility

of atrial fibrillation in patients undergoing ablation for atrial flutter. Sparks et al. (2000)

studied patients with paroxysmal atrial flutter and observed a reduction in ERP

immediately following 5‐10 minutes of induced atrial flutter with a return to baseline

ERP over a period of minutes. In patients with chronic atrial flutter however, ERP

remained low out to 30 minutes but was seen to recover (in addition to sinus node

function) when retested at 3 weeks. Further evidence for remodelling of the sinus

node with atrial flutter is reported by Daoud et al. (2002) who studied patients with

chronic atrial flutter following ablation and observed impaired sinus node function

which recovered over a 3 month period.

These reports of electrical remodelling as a result of atrial flutter characterised by ERP

shortening and impairment of sinus node function have underlined the importance of

recent arrhythmia in generating these features. Whether such features persist when

tested remote from an episode of atrial flutter is critical when elucidating the features

of the underlying atrial substrate (rather than the effects of rate related remodelling)

that may aid the progression of arrhythmia to atrial fibrillation.

4.4.4 Substrate in Conditions Predisposed to Atrial Arrhythmia

While studies have implicated the role of shortened ERP resulting from arrhythmia

itself in the development and maintenance of atrial arrhythmia (Morillo et al. 1995;

111

Wijffels et al. 1995), evaluation of conditions predisposed to atrial flutter and atrial

fibrillation (but without arrhythmia) have provided evidence to the contrary. Li et al.

(1999) developed a canine model of congestive heart failure and demonstrated "atrial

remodelling of a different sort" to that seen due to atrial fibrillation alone. An increase

in ERP at short cycle lengths associated with an increase in the heterogeneity of

conduction resulted in longer durations of atrial fibrillation despite the increase in

refractoriness. Verheule et al. (2003) observed an increase in ERP but a propensity for

atrial fibrillation in a canine model of mitral regurgitation, presumed to be due to

abnormalities in conduction and structural remodelling. In humans, patients with atrial

septal defects have demonstrated prolonged ERP and delayed conduction across the

crista terminalis compared to reference patients (Morton et al. 2003). Patients with

congestive heart failure similarly have significant conduction abnormalities and sinus

node dysfunction in association with prolonged ERP (Sanders et al. 2003). Kistler et al.

(2004) have demonstrated widespread and site‐specific conduction abnormalities in a

progressive manner with increasing age. Similar changes have been documented by

Sanders et al. (2004a) and John et al. (online early) in patients with sinus node disease

and rheumatic mitral stenosis, respectively.

These experimental and clinical studies provide compelling evidence that the

predominant electrophysiological features of the substrate for atrial arrhythmias,

excluding the effect of arrhythmia itself, are related to conduction abnormalities

rather than changes in refractoriness. In the current study we have shown that

patients with atrial flutter, studied remote from arrhythmia, have right atrial

conduction abnormalities and sinus node dysfunction, without obvious reduction in

112

refractoriness. These findings are consistent with the studies to have evaluated the

substrate predisposing to atrial fibrillation.

4.4.5 Implications

This study has confirmed the existence of significant widespread remodelling of the

atria in patients with atrial flutter remote from an arrhythmic event characterised by

increased linear conduction times, sinus node dysfunction and site‐specific conduction

abnormalities. Patients were studied remote in time from the arrhythmia itself to

evaluate the underlying substrate without the known confounding effects of

arrhythmia itself, which may well further perpetuate atrial arrhythmia. These findings

strongly argue that patients with atrial flutter have a pre‐existing diffuse substrate

which may contribute to both the frequent co‐existence of atrial fibrillation and the

development of atrial fibrillation after ablation for atrial flutter.

4.4.6 Limitations

While the abnormalities observed in this study may constitute the substrate

predisposing to atrial fibrillation, the development of clinical atrial fibrillation is

complex and depends not only on substrate but also on other factors such as triggers

and perpetuators that were not addressed by this study (Allessie et al. 2001). Access to

the left atrium was not mandated by the clinical procedure therefore no left atrial data

could be collected. Finally, although no patient had symptoms for a month and

patients were monitored for a week prior to the procedure to ensure our evaluation

113

was remote from an episode of atrial flutter, we cannot exclude an additive effect of

rate‐related remodelling, although there is a substantial amount of evidence that

electrical remodelling reverses within a short time after restoration of atrial flutter to

sinus rhythm (Sparks et al. 2000).

4.5 Conclusion

Patients with atrial flutter, remote from an arrhythmic episode, demonstrate

conduction abnormalities and altered sinus node function, without reduction in atrial

refractoriness. These findings suggest that the cause of the observed close inter‐

relationship of atrial flutter and atrial fibrillation may be reflected in this atrial

substrate.

114

Table 4.1


Atrial Flutter

(n=8)

Reference

(n=12)

p value

Age (years) 66±8 59±12 0.2

Structural heart disease, n (%) 2 (25) 1 (8) 0.5


Left atrial area (cm2) 22±5 19±4 0.3

Right atrial area (cm2) 19±3 13±6 0.1

LV end‐diastolic dimension (mm) 49±4 46±6 0.3

LV end‐systolic dimension (mm) 30±7 30±5 0.9


LV = left ventricular.

115

Figure 4‐1


Left anterior oblique view of a decapolar catheter in the coronary sinus, a

duodecapolar catheter along the crista terminalis with its stabilising sheath and a

duodecapolar catheter along the lateral right atrium. Subsequently, the duodecapolar

catheter was positioned at the high‐septal right atrium for ERP testing (not shown).

116

Figure 4‐2

Regional effective refractoriness

Mean (+ standard deviation) effective refractory period at the five sites tested

following a 600ms drive train (top) and a 450ms drive train (bottom) in patients with

atrial flutter and reference patients. All values for 600ms are higher than the same

patient group and site value for 450ms.

*p<0.05. dCS – distal coronary sinus. pCS – proximal coronary sinus. LLRA – low lateral

right atrium. HLRA – high lateral right atrium. HRAS – high right atrial septum.

117

Figure 4‐3


Mean (+ stand

ard de

viation) num

ber (le

ft) and maxim

um duration (right) of dou

ble po

tentials or fractio

nated

signals alon

g the crista terminalis during pacing

(S1) at 600ms (top

) and 450m

s (bottom) and with

the

earliest

captured

extra‐stim

ulus (S2) at the proxim

al coron

ary sinu

s (pCS) a

nd low‐la

teral right atrium (LLRA) in patie

nts with

atrial flutter a

nd referen

ce patients. p<0.03 for all com

parisons between atrial flutter a

nd referen

ce patients.

118

Chapter 5. Right atrial electroanatomical substrate of atrial flutter

5.1 Introduction

The underlying substrate of typical atrial flutter is not fully understood. Atrial flutter

has a close inter‐relationship with atrial fibrillation and the two rhythms frequently co‐

exist (Halligan et al. 2004). Ablation of atrial flutter has been observed to significantly

reduce the occurrence of atrial fibrillation in the short term (Schmieder et al. 2003).

Despite this reduction a significant proportion, perhaps even the majority with long‐

term follow‐up, eventually develop atrial fibrillation (Paydak et al. 1998; Chinitz et al.

2007; Ellis et al. 2007; Meissner et al. 2007; Moreira et al. 2008). Key to the

understanding of the close inter‐relationship for atrial flutter and atrial fibrillation is

the recognition that typical flutter almost always develops from antecedent atrial

fibrillation of variable duration (Shimizu et al. 1991; Waldo and Cooper 1996;

Roithinger et al. 1997). Emerging evidence implicates abnormal atrial structural

remodelling and its resultant electrophysiological milieu in clinical conditions known to

be predisposed to atrial fibrillation (Li et al. 1999; Morton et al. 2003; Sanders et al.

2003; Verheule et al. 2003; Kistler et al. 2004; Sanders et al. 2004a; Kistler et al. 2006;

John et al. online early). Whether patients with atrial flutter also demonstrate

abnormal substrate outside of the well characterised circuit of typical flutter is not

known.

Critical to our management of patients with both atrial flutter and atrial fibrillation is

an understanding of the interactions and shared characteristics of these rhythms.

Importantly, electrical characteristics of patients with atrial arrhythmia are known to

119

be affected by recent arrhythmia (Wijffels et al. 1995). We hypothesised that patients

with atrial flutter would demonstrate an abnormal atrial substrate to account for the

frequent development of atrial fibrillation. In order to evaluate this we performed

high‐density electroanatomical mapping of the right atrium to assess voltage and

conduction characteristics of the atrial flutter substrate, remote from recent

arrhythmia.

5.2 Methods


This study comprised 12 patients without antecedent atrial fibrillation undergoing

ablation for typical atrial flutter and a reference group of 9 patients undergoing

radiofrequency ablation for atrioventricular reentry tachycardia or atrioventricular

nodal reentry tachycardia. Patients with atrial flutter were selected on the basis of

having ≥2 previous episodes of typical atrial flutter confirmed on a 12‐lead

electrocardiogram. These patients were documented to be in sinus rhythm and

described no symptoms attributable to atrial flutter for at least 1 month prior to

ablation. In addition, all patients underwent continuous monitoring for 7 days prior to

the study and were excluded if they had evidence of any atrial arrhythmia >30

seconds.



120


Adelaide Hospital.



utilising midazolam and fentanyl. In patients with atrial flutter the study protocol was

performed before ablation, while in the reference group this was done after ablation.

Following the study protocol, patients with atrial flutter underwent cavotricuspid

isthmus ablation with a delivered power of 35W with irrigation rates of 30‐60mL/min

and an endpoint of bi‐directional isthmus block.

5.2.3 Electroanatomical Study Protocol

Electroanatomical maps were created of the right atrium during sinus rhythm using the

CARTO mapping system (Biosense‐Webster). The electroanatomical mapping system

has been previously described in detail; the accuracy of the sensor position has been

previously validated and is 0.8mm and 5 degrees (Gepstein et al. 1997). In brief, the

system records the surface ECG and bipolar electrograms filtered at 30‐400Hz from the

mapping and reference catheters. Endocardial contact during point acquisition was

facilitated by electrogram stability, fluoroscopy and the catheter icon on the CARTO

system. Points were acquired in the auto‐freeze mode if the stability criteria in space

(≤6mm) and local activation time (≤5ms) were met. Mapping was performed with an

equal distribution of points using a fill‐threshold of 15mm. Editing of points was

121

performed off‐line. Local activation time was manually annotated to the peak of the

largest amplitude deflection on bipolar electrograms; in the presence of double

potentials this was annotated at the largest potential. If the bipolar electrogram

displayed equivalent maximum positive and negative deflections, the maximum

negative deflection on the simultaneously acquired unipolar electrogram was used to

annotate the local activation time. Points not conforming to the surface ECG P wave

morphology or <75% of the maximum voltage of the preceding electrogram were

excluded. Regional atrial bipolar voltage and conduction velocity were analysed as

previously described and are detailed below (Sanders et al. 2003; Kistler et al. 2004;

Sanders et al. 2004a; John et al. online early).

5.2.3.1 Voltage Analysis

For the purposes of evaluating regional voltage differences, each atrium was

segmented using previously validated offline software (Kuklik et al. 2004). The right

atrium was segmented as the high‐ and low‐lateral right atrium, high‐ and low‐

posterior right atrium, high‐ and low‐septal right atrium, and anterior right atrium. At

each of these regions an average voltage of 10 points was determined. The voltage of

points identified by region was exported for analysis. Low voltage points were defined

as points with a bipolar voltage ≤0.5mV and electrically silent points as the absence of

recordable activity or a bipolar voltage amplitude ≤0.05mV (the noise level of the

system).

5.2.3.2 Conduction Velocity Analysis

Isochronal activation maps (5ms intervals) of the atria were created and regional

conduction velocity determined in the direction of the wave‐front propagation (least

122

isochronal crowding). An approximation of conduction velocity was determined by

expressing the distance between two points as a function of the difference in local

activation time. Mean conduction velocity for each region was determined by

averaging the conduction velocity between 5 pairs of points, as previously described

(Sanders et al. 2003; Kistler et al. 2004; Sanders et al. 2004a; John et al. online early).

For the purposes of evaluating regional conduction differences, each atrium was

segmented as above.

5.2.3.3 Complex Electrograms

The proportion of points demonstrating complex electrograms was determined using

the following definitions: (i) Fractionated signals – complex activity of ≥50ms duration;

and (ii) Double potentials – potentials separated by an isoelectric interval where total

electrogram duration was ≥50ms.





between means were analysed using analysis of variance and paired or unpaired t tests

as appropriate. Comparisons with adjustment for multiple sampling within patients

were performed using a Mixed Linear Model for continuous data or a logistic

Generalised Estimating Equation for categorical data. Statistical tests were performed

using SPSS 15 (SPSS Inc) and statistical significance was set at p<0.05.

123

5.3 Results


Baseline patient characteristics are summarised in Table 5.1. Patients with atrial flutter

had a median atrial flutter history of 16 months (IQR 8‐22), however they had no

history of any arrhythmia for at least 1 month (range 1‐7 months) prior to the study.

Groups were well matched for age, left ventricular dimensions and hypertrophy,

however patients with atrial flutter had significantly larger left atria on trans‐thoracic

echocardiogram.

A total of 97±19 points per patient were analysed in the right atrium using

electroanatomical mapping.

5.3.2 Structural and Voltage Abnormalities

Mean right atrial volume by electroanatomical map was significantly greater in

patients with atrial flutter compared to reference patients (121±30 versus 83±24mL,

p=0.005). The mean right atrial bipolar voltage was reduced in patients with atrial

flutter compared to reference patients (2.1±0.5 versus 3.0±0.5mV, p<0.001). Regional

differences in bipolar voltage are illustrated in Figure 5‐1; significant differences

between patients with atrial flutter and reference patients were seen in high‐lateral,

high‐posterior and anterior regions. Additionally, points at the high‐posterior region

were more likely to be low voltage (≤0.5mV) in patients with atrial flutter (odds ratio

3.7, 95% CI 2.1‐6.7). No significant areas of electrical silence (≤0.05mV) were observed.

124

Illustrative examples of electroanatomical maps in representative patients are shown

in Figure 5‐2.

5.3.3 Abnormalities in Conduction Velocity

Patients with atrial flutter had slower mean conduction velocity during sinus rhythm

compared to reference patients (1.2±0.2 versus 2.1±0.6mm/ms, p<0.001). Figure 5‐3

illustrates regional conduction velocities, which were significantly slower in patients

with atrial flutter compared to reference patients at the high‐lateral, low‐posterior,

low‐septal and anterior right atrium.

The right atrial activation time was significantly prolonged in patients with atrial flutter

compared to reference patients (107±23 versus 85±14ms, p=0.02).

5.3.4 Complex Electrograms

Patients with atrial flutter demonstrated a significantly greater number of points with

double potentials or fractionated signals than reference patients (16±4 versus 10±6%,

p=0.006). These points were distributed throughout both atria with a clustering at the

high‐posterior, high‐lateral and high‐septal regions of the right atrium (Figure 5‐2).

125

5.4 Discussion


This study utilised electroanatomical mapping to demonstrate new information on the

nature of atrial abnormalities in patients with typical atrial flutter, remote from an

arrhythmic event. The major findings are as follows:

(i) Structural abnormalities characterised by atrial dilatation and lower mean

right atrial voltage suggesting the loss of atrial myocardium.

(ii) Slower conduction velocities.

(iii) An increase in the proportion of electrograms characterised as double

potentials or fractionated signals.

Thus, the present study suggests that patients with atrial flutter, studied remote from

an arrhythmic event, have widespread and persistent abnormal atrial substrate. We

posit that these diffuse atrial abnormalities are partly responsible for the frequent

development of atrial fibrillation in patients with atrial flutter.

5.4.2 Substrate in Clinical Conditions Associated with Atrial Arrhythmia

Sanders et al. have demonstrated significant electroanatomical remodelling with areas

of low voltage and electrical scar with associated widespread conduction abnormalities

in patients with congestive heart failure (2003) and sinus node disease (2004a)

compared to control groups. Kistler et al. (2004) have demonstrated by

electroanatomical mapping studies that ageing is associated with progressive regional

126

conduction slowing and structural changes that include areas of low voltage. Perhaps

the most marked changes have been shown in patients with rheumatic mitral stenosis

in whom striking differences in voltage, conduction velocity and the proportion of

complex electrograms are seen when compared to controls (John et al. online early). In

addition to the clinical studies, animal models of such diseases have revealed similar

features in the abnormal substrate contributing to the potential for arrhythmia (Li et

al. 1999; Verheule et al. 2003; Kistler et al. 2006).

These studies provide compelling evidence that the predominant contributors to the

substrate for atrial arrhythmias are diffuse structural and conduction abnormalities. In

the current study we have shown that patients with atrial flutter, studied remote from

arrhythmia, have right atrial structural remodelling with associated conduction

abnormalities. These findings are thus consistent with the studies to have evaluated

the clinical substrate predisposing to atrial fibrillation.

5.4.3 Substrate in Experimental Models of Atrial Flutter

Early work to create atrial flutter in animals included a crush injury model by creating a

lesion between the venae cavae (Rosenbleuth and Garcia‐Ramos 1947). Based on the

epicardial maps, the authors deduced that the reentry loop circled around the atrial

crush lesion. Much work on atrial flutter has been performed in the sterile pericarditis

dog model which has been shown to dependably produce this arrhythmia to allow

characterisation of the circuit (Page et al. 1986). One of the critical features

demonstrated by this model is that a sufficient length of functional block is critical to

the maintenance of atrial flutter (Ortiz et al. 1994). Interestingly, the time course of

127

atrial flutter developing is similar to that of the post‐bypass arrhythmias seen clinically

(Waldo and Cooper 1996); this underscores the role of pericardial inflammation in the

genesis of arrhythmia in cardiac surgery patients. More recently, the role of

inflammation in atrial flutter has been highlighted in a study in which the usually

reliable induction of atrial flutter in the sterile pericarditis model was abolished by pre‐

treatment with prednisone (Goldstein et al. 2008). Histological and gross inspection of

the atria revealed marked structural changes of inflammation only in the untreated

group, all of whom (bar one with atrial fibrillation) developed atrial flutter as expected.

A sheep model of atrial flutter demonstrated that sustained atrial flutter for 4 weeks

allows the development of atrial substrate such that conversion to atrial fibrillation is

achieved by extra‐stimuli and that this new rhythm is able to be maintained (Morton

et al. 2002).

5.4.4 Substrate in Clinical Atrial Flutter

Studies of human atrial flutter have been mainly concerned with elucidating the typical

flutter circuit within the right atrium. This has now been well characterised with the

posterior and anterior barriers to the circuit defined as the crista terminalis and the

tricuspid valve, respectively (Olgin et al. 1995; Kalman et al. 1996). Differences in the

atrial flutter circuit between young and old have been published, with younger

patients tending toward slowest conduction in the lateral isthmus while older patients

demonstrate a low voltage zone and slowest conduction in the medial isthmus (Huang

et al. 2008). The substrate of patients with both atrial fibrillation and atrial flutter may

differ from that seen in atrial fibrillation alone. Patients with typical atrial flutter in

128

addition to paroxysmal atrial fibrillation have been shown to have more frequent

recurrence of arrhythmia following pulmonary vein isolation with accompanying

isthmus line (Moreira et al. 2007). This suggests that non‐pulmonary vein sites are

more important to the maintenance of atrial fibrillation in patients with co‐existent

atrial flutter and supports the view that the atrial substrate in atrial flutter is indeed

diffuse throughout the atria.

5.4.5 Implications

This study has confirmed the existence of significant diffuse remodelling of the atria in

patients with atrial flutter remote from an arrhythmic event characterised by regional

slowing in conduction and structural change reflected in lower voltage, larger atria and

widespread conduction abnormalities. Patients were studied remote in time from the

arrhythmia itself to evaluate the underlying substrate without the known confounding

effects of arrhythmia itself, which may well further perpetuate atrial arrhythmia. These

findings strongly argue that patients with atrial flutter have a pre‐existing diffuse

substrate which may contribute to both the frequent co‐existence of atrial fibrillation

and the development of atrial fibrillation after ablation for atrial flutter.

5.4.6 Limitations

Access to the left atrium was not mandated by the clinical procedure therefore no left

atrial data could be collected. While the abnormalities observed in this study may

constitute the substrate predisposing to atrial fibrillation, the development of clinical

129

atrial fibrillation is complex and depends not only on substrate but also on other

factors such as triggers and perpetuators that were not addressed by this study

(Allessie et al. 2001).

5.5 Conclusion

Patients with atrial flutter, remote from an arrhythmic episode, demonstrate structural

abnormalities characterised by loss of myocardial voltage, conduction slowing and an

increase in the proportion of abnormal electrograms. These findings suggest that the

cause of the observed close inter‐relationship of atrial flutter and atrial fibrillation may

be reflected in this atrial substrate.

130

Table 5.1


Atrial Flutter

(n=12)

Reference

(n=9)

p value

Age (years) 55±10 59±15 0.5





LV end‐diastolic dimension (mm) 52±4 50±6 0.5

LV end‐systolic dimension (mm) 30±6 34±5 0.2


LV = left ventricular.

131

Figure 5‐1

Regional bipolar voltage

Mean (+ standard deviation) bipolar voltage of the 7 right atrial regions from the

electroanatomical map.

*p<0.05. H – high. L – low.

132

Figure 5‐2

Example electroanatomical bipolar voltage maps

Representative CARTO maps of a patient with atrial flutter (right) and a reference

patient (left). Both are oriented in the posterior‐anterior projection and are of the

same scale. The colour scale is identical in both images with red representing voltage

≤0.5mV and purple being voltage ≥5mV. In addition to having greater regions of low

voltage (red), the patient with atrial flutter has evidence of conduction abnormalities

in the form of fractionated signals (pink tags) and double potentials (blue tags).

133

Figure 5‐3

Regional conduction velocity

Mean (+ standard deviation) conduction velocity in the 7 right atrial regions from the

electroanatomical map.

*p<0.05. H – high. L – low.

134

Chapter 6. High‐density mapping of the sinus node to characterise the

remodelling resulting from chronic atrial flutter

6.1 Introduction

The anatomical sinus node is a small localised structure that is located at the junction

of the superior vena cava with the right atrium (James 1977; Anderson et al. 1979). In

contrast, studies of the canine functional sinus node complex have identified that the

area of initial cardiac depolarisation exceeds that of the anatomical sinus node by 3 to

4‐fold (Boineau et al. 1980). Although this implicates the existence of widely

distributed nodal cells capable of automaticity, histological sectioning of the region by

a range of investigators has failed to identify islands of nodal cells separate from the

sinus node itself (James 1977; Anderson et al. 1979; Boineau et al. 1980; Sanchez‐

Quintana et al. 2005). Boineau et al. (1988) undertook several of the early seminal

evaluations of the sinus node and have proposed two alternative hypotheses for extra‐

nodal sinus complex origins; (i) the existence of specialised conduction tracts from the

anatomically fixed nodal source to remote exit sites, and (ii) a coordinated exchange of

dominance amongst competing pacemakers via autonomic modulation. For the former

explanation, the presence of nodal radiations into adjacent atrial myocardium through

electrically isolating connective tissue supports this theory, however no insulated

rapidly‐conducting tissue analogous to that found in the ventricles has been

conclusively demonstrated (Anderson et al. 1981a). Supporting the latter hypothesis,

autonomically‐mediated shifts in activation sites appear linked to changes in heart rate

(Littmann et al. 1990), however the mechanism by which a stable heart rate is

maintained without faster pacemakers dominating is yet to be demonstrated.

135

The observations above were based on conventional mapping techniques. However,

newer mapping technology has provided novel insights into arrhythmia mechanisms

not previously afforded using conventional techniques. Higa et al. (2004) used non‐

contact mapping, and demonstrated “preferential pathways” from earliest activation

to the site where break‐out occurs to the remaining atria in ectopic atrial tachycardia.

Whether similar pathways exist to account for the discrepancy between the

anatomical and functional sinus node complex is not known.

This study aimed to characterise the electrical properties of the functional sinus node

complex in the normal heart. To further explore the sinus node complex activity, the

same characterisation was performed in a group known to have impaired sinus node

function (Sparks et al. 2000; Daoud et al. 2002) to establish how sinus node activity is

affected by atrial remodelling. The anatomical and electrical resolution provided by the

three‐dimensional, simultaneous, high‐density EnSite Array mapping system (St Jude

Medical) provides the ability to examine the extent and variability of endocardial sinus

activation with detail not previously available.

6.2 Methods


Fifteen patients with structurally normal hearts undergoing radiofrequency ablation

for atrioventricular nodal reentry tachycardia were studied as a reference group. To

evaluate the effects of remodelling, a further 16 patients undergoing ablation for

chronic typical atrial flutter (median 8 months, IQR 6‐23) were studied. The latter

136

group was selected on the basis of the well documented occurrence of sinus node

dysfunction with atrial flutter in previous studies (Sparks et al. 2000; Daoud et al.

2002).

All anti‐arrhythmic medication including beta‐blockers and calcium channel blockers,

were ceased ≥5 half‐lives prior to the study. No patient had received amiodarone in

the preceding 6 months. All patients provided written informed consent to the study

protocol that was approved by the Human Research Ethics Committee of the Royal

Adelaide Hospital.



utilising midazolam and fentanyl. After confirming the arrhythmia mechanism,

reference group patients underwent slow pathway ablation and patients with flutter

underwent cavotricuspid isthmus ablation with an end‐point of bi‐directional block

confirmed by activation mapping. The study protocol was conducted immediately

following ablation for the clinical arrhythmia.

The following catheters were used for the study protocol: (i) a 10‐pole catheter (2‐5‐

2mm inter‐electrode spacing, Daig Electrophysiology) positioned within the coronary

sinus with the proximal bipole at the coronary sinus ostium as determined in the best

septal left anterior oblique position; (ii) a 20‐pole “crista” catheter (1‐3‐1mm inter‐

electrode spacing, Biosense‐Webster) placed with the distal tip superiorly such that

the second bipole lay at the superior vena cava‐right atrial junction, stabilised by a long

sheath (CSTA, Daig Electrophysiology) to ensure close apposition along the crista

137

terminalis; (iii) a roving 4mm ablation catheter (Celsius, Biosense‐Webster); and (iv) a

multi‐electrode array (EnSite Array, St Jude Medical) mounted on a 7.5mL balloon

positioned in the superior right atrium (Figure 6‐1). Maps were created of the right

atrial geometry using the roving catheter to trace the endocardial surface while the

EnSite system recorded its three‐dimensional position, as previously described

(Schilling et al. 1998). Intra‐cardiac echocardiography (Ultra ICE, Boston Scientific) was

used to standardise catheter positioning and also to localise the crista terminalis and

superior vena cava‐right atrial junction (Figure 6‐2).

6.2.3 Conventional Evaluation of Sinus Node Function

6.2.3.1 Baseline Sinus Cycle Length

The baseline sinus cycle length was determined over 10 consecutive sinus cycles after

10 minutes of rest.

6.2.3.2 Sino‐atrial Conduction Time

The SACT was determined after an 8‐beat pacing train according to the formula: SACT

= (return ‐ basic cycle length)/2 (Narula et al. 1978). SACT was determined 3 times and

averaged.

6.2.3.3 Corrected Sinus Node Recovery Time

The CSNRT was determined after a 30‐second drive train at cycle lengths of 600 and

450ms, correcting for the baseline cycle length. At each cycle length, CSNRT was

determined 3 times and averaged.

138

6.2.3.4 Conduction Properties at the Crista Terminalis

Site‐specific conduction properties at the crista terminalis were quantified by

recording the presence of electrograms with discrete double potentials separated by

an isoelectric interval or complex fractionated activity of ≥50ms duration on each

recording bipole of the crista terminalis catheter. Both the number of bipoles

demonstrating these conduction characteristics and the maximum electrogram

duration were determined over ten beats in sinus rhythm and one beat immediately

after the cessation of each pacing train.

6.2.4 High‐Density Simultaneous Unipolar Mapping of Sinus Node Function

Mapping of the sinus node complex was performed during baseline sinus rhythm and

also after pace suppression of sinus pacemaker activity following 30‐second drive

trains at 600 and 450ms. Sinus node activation was characterised by determining the

positions of earliest activation (EA) and sinus break‐out (SBO). EA was defined as the

earliest unipolar electrogram displaying a “QS” pattern and SBO as the earliest

unipolar electrogram displaying an “RS” pattern and a sudden increase in dV/dt

indicating rapid depolarisation of surrounding myocardium, as previously described

(Higa et al. 2004). For review of the recorded data the high‐pass filter was set to 1Hz to

allow for slow conduction of the activation wave‐front. The position of the virtual

electrodes in the region of interest was repeatedly adjusted to find the points of

interest. The following parameters were evaluated from simultaneous unipolar

recordings of right atrial activation from the balloon array.

139

6.2.4.1 Location of Points of Earliest Activation and Sinus Break‐out

The location of the EA and SBO were determined to define the position of sinus origin

relative to the superior vena cava‐right atrial junction and the crista terminalis during

baseline sinus rhythm and following pace suppression of the sinus node at cycle

lengths of 600 and 450ms on three occasions each. The superior vena cava‐right atrial

junction was defined as the lateral confluence of the superior vena cava, right atrium

and right atrial appendage, as previously described (Callans et al. 1999). This was

confirmed using intra‐cardiac echocardiography demonstrating, at this level, the

superior lateral crista terminalis in proximity with the orifice of the right atrial

appendage (Figure 6‐2). The crista terminalis was then marked along its length with a

vertical reference line from the superior vena cava‐right atrial junction to the inferior

vena cava. The parallel distance from the origin of this line to the point of interest was

recorded, as was the perpendicular distance (Figure 6‐2). This gave a position in both

the supero‐inferior and antero‐posterior dimensions. In the rare case of a multi‐centric

SBO point (Movie 6.1), the more cranial was measured.

6.2.4.2 Preferential Conduction Pathways

The path between the EA and SBO was defined as the route of preferential conduction

(Figure 6‐3), as previously described (Higa et al. 2004). We determined the time to

conduct between EA and SBO, and the direction taken. The activation pattern from

SBO to the right atrial in relation to the crista terminalis, including transverse

conduction through the crista, was observed and atrial propagation qualitatively

assessed. Additionally, right atrial activation time from EA to final discernable right

atrial activation as recorded by the balloon array was determined.

140

6.2.4.3 Voltage in the Right Atrium and Crista Terminalis

Unipolar voltage at the points of EA and SBO was recorded. Electrograms for 100ms

following EA were exported from the EnSite Array for further analysis. This included

256 evenly distributed electrograms throughout the right atrium and 10 evenly spaced

electrograms along the length of the crista terminalis. From these exports, peak

negative unipolar voltage of the entire right atrium and along the crista terminalis was

determined. Additionally, right atrial surface area and distance to the array was able to

be determined from the exported geometry.



quartile range as appropriate to its distribution. Categorical variables are reported as

number and percentage. Comparisons between groups were performed using

unpaired t tests or Mann‐Whitney U tests. Proportions were compared using Fisher’s

Exact test. Comparisons with adjustment for multiple sampling within patients were

performed using a Mixed Linear Model for continuous data or a logistic Generalised

Estimating Equation for categorical data. Statistical tests were performed using SPSS

15 (SPSS Inc) and statistical significance was set at p<0.05.

141

6.3 Results


Baseline patient characteristics are summarised in Table 6.1. Mean distance from the

array to the right atrial geometry was 30±3mm and the crista terminalis was 23±3mm,

well within the reported limits of accuracy (Thiagalingam et al. 2004b). Baseline sinus

cycle length did not differ significantly between reference patients and those recently

ceasing chronic flutter (775±104ms versus 784±193ms, p=0.9). Differences were seen

between the reference group and patients with flutter in right atrial surface area

(132±12 versus 163±34cm2, p=0.003), P wave duration (109±13 versus 147±13ms,

p<0.001), inter‐caval distance (70±7 versus 87±8mm, p<0.001), and right atrial

activation time (92±12 versus 120±15ms, p<0.001).

6.3.2 Sinus Node Dysfunction and Atrial Remodelling

Reference patients had normal measures of sinus node function, including SACT

(83±33ms) and CSNRT (at 600ms: 245±88ms; at 450ms: 253±96ms). Patients with

flutter demonstrated longer SACT (150±100ms, p=0.03) and longer CSNRT (at 600ms:

339±151ms, p=0.04; at 450ms: 498±409ms, p=0.03). Site‐specific conduction

abnormalities at the crista terminalis during sinus rhythm were significantly less in the

reference group than in patients with flutter (Figure 6‐4): the number of double

potentials or fractionated electrograms was 2.1±1.8 versus 5.4±2.9, p=0.002; the

maximum duration of such signals was 60±15 versus 72±14ms, p=0.04. These

142

differences were even more marked during and immediately following 30‐second

pacing trains (Table 6.2).

6.3.3 Extent and Shift of the Atrial Pacemaker Complex

As determined by the multi‐polar array, the mean distance from the superior vena

cava‐right atrial junction to the point of EA was 4±4mm and SBO 9±6mm in reference

patients. The mean distance from the crista terminalis in an anterior direction was

8±11mm for EA and 14±8mm for SBO. For patients with flutter, the distance inferiorly

from the superior vena cava‐right atrial junction was greater (EA: 15±12mm, p=0.003;

SBO: 23±11mm, p<0.001). The distance from the crista terminalis anteriorly to the EA

(5±14mm, p=0.5) or SBO (16±7mm, p=0.5) did not differ for patients with flutter.

Normalised to mean inter‐caval distance for this group with larger right atria, the

distance inferiorly to EA (p=0.007) and SBO (p=0.003) remained significantly different

from the reference group.

Following 30‐second pacing trains the mean distance from the superior vena cava‐right

atrial junction to EA and SBO points changed significantly for reference patients (EA

4±4 to 13±8mm, p=0.006; SBO 9±6 to 16±10mm, p=0.02); representing a mean caudal

shift for EA of 10±9mm and for SBO of 7±8mm. For patients with flutter the mean

distance from the superior vena cava‐right atrial junction to EA and SBO points also

changed significantly (EA 15±12 to 19±12mm, p=0.045; SBO 23±11 to 28±9mm,

p=0.02). However, caudal shift was markedly diminished in patients with flutter

compared to the reference group; EA was 4±7mm and SBO 4±6mm. The change in

distance from the crista terminalis anteriorly to the EA or SBO following pacing did not

143

significantly differ for either patient group. Figure 6‐5 demonstrates the mean location

and caudal shift of the physiological pacemaker complex and a representative example

is shown in Figure 6‐6. The median number of beats for activation to resume baseline

position was 2 (IQR 1 to 4, range 1 to 28 beats) and did not significantly differ between

patient groups.

6.3.4 Preferential Pathway Conduction

The mean time to conduct along the preferential pathway from EA to SBO was 15±5ms

for the reference group. Patients with flutter took significantly longer (23±8ms,

p=0.005) to conduct between the two points. While the SBO was located at or anterior

to the crista terminalis in all patients, the preferential pathway between the EA to the

SBO was directed inferiorly in 63%, anteriorly in 22%, posteriorly in 10% and superiorly

in 5% of beats analysed; no difference was seen between patient groups.

6.3.5 Beat‐to‐beat Variation in Sinus Node Function

The beat‐to‐beat site of EA ranged within an individual from 0 to 41mm (median 19,

IQR 7‐31) from the superior vena cava‐right atrial junction suggesting the capacity for

multi‐centric activation. We did not observe simultaneously occurring multi‐centric EA

points. Beat‐to‐beat variation of SBO ranged within an individual from 0 to 33mm

(median 12, IQR 8‐23). In contrast to EA, there were 3 controls who demonstrated dual

near‐simultaneous SBO points during a single beat (Movie 6.1). In all 3 instances, one

of the SBO points was on the crista terminalis.

144

6.3.6 Crista Terminalis Conduction

More frequent transverse conduction through the crista terminalis was seen during

sinus rhythm beats in the reference group than in the patients with flutter (38% versus

4%, p=0.003), almost invariably through the superior third of the crista near the level

of the SBO point. For those beats analysed in which there was no conduction through

the crista terminalis, the delay between anterior and posterior activation of the crista

was 17±6ms for reference patients and 24±7ms for patients with flutter (p=0.02). The

circuitous activation pattern subsequent to SBO in such patients was in 2 directions:

inferiorly down the anterior border of the crista terminalis to the anterior right atrium,

and supero‐posteriorly over the basal appendage and then inferiorly down the

posterior right atrium (Figure 6‐7).

6.3.7 Voltage Findings

For reference patients, mean unipolar voltage was 1.8±0.6mV for the right atrium,

1.9±0.6mV along the crista terminalis, 0.12±0.10mV at the point of EA and 3.2±1.5mV

at the point of SBO. The average maximum unipolar voltage recorded was 5.0±1.8mV

for the right atrium and 3.5±1.4mV along the crista terminalis. Patients with flutter

trended to lower mean right atrial voltage (1.4±0.6mV, p=0.1), while mean voltage at

the crista (1.3±0.6mV, p=0.009) and point of SBO (1.5±1.4mV, p=0.005) was

significantly lower compared to reference patients. Voltage at the point of EA did not

differ (0.12±0.13mV, p=0.9). As with mean voltage, maximum voltage for the right

atrium trended to lower for patients with flutter (3.9+2.1mV, p=0.2) and was

145

significantly less at the crista terminalis (2.2±1.2mV, p=0.01). Differences in voltage

parameters are shown in Table 6.3.

6.4 Discussion


This study utilised high‐density simultaneous endocardial unipolar mapping of sinus

rhythm to describe the characteristics of the functional sinus node complex and the

effects of remodelling due to chronic atrial flutter. The major findings are as follows:

(i) The sinus node complex in normal hearts displays a dynamic range of activation

sites along the postero‐lateral right atrium.

(ii) Preferential pathways of conduction exist between the sinus node and the exit

of sinus activity to the atria.

(iii) There were multiple origins of sinus activation (EA) and exit sites to the atria

(SBO), providing evidence of multi‐centricity of the sinus node complex.

(iv) Remodelled atria demonstrate structural change reflected by loss of voltage

that was associated with more caudal activation, slower conduction time along

preferential pathways, only modest shifts of the functional pacemaker complex

and more frequent conduction block across the crista terminalis with resultant

circuitous conduction of the sinus impulse.

146

These findings extend our knowledge of the function of the highly complex human

sinus node by demonstrating the presence of preferential pathways of conduction

within the functional sinus node complex.

6.4.2 The Sinus Node: Anatomical and Functional Considerations

The sinus node consists of specialised nodal cells lying immediately sub‐epicardially at

the superior pole of the sulcus terminalis of the right atrium and has a mean length of

13.5mm (range 8‐21.5mm) (Sanchez‐Quintana et al. 2005). The variability of the

functional sinus node position exceeds that of the histologically defined sinus node by

a factor of three to four (Boineau et al. 1980). Shifts in the origin of the sinus node

impulse along the sulcus terminalis under autonomic influences have been observed

by several investigators since early last century (Lewis et al. 1910; Geesbreght and

Randall 1971).

Further evidence of a widely distributed functional pacemaker complex comes from

clinical studies that have aimed to modify sinus activity in patients with inappropriate

sinus tachycardia. In contrast to the single lesions effective for ectopic atrial

tachycardia treatment, these procedures require extensive ablation along the length of

the crista terminalis to achieve functional inhibition of the sinus node (Lee et al. 1995).

Boineau et al. (1978; 1980; 1984) performed detailed epicardial mapping of animal

hearts and observed that the sinus P wave may arise from a “pacemaker complex”,

distributed over an extensive extra‐nodal area spanning the superior vena cava‐right

atrial junction to the inferior vena cava. A strong correlation between heart rate and

functional sinus origin was observed with vagal stimulation and isoprenaline infusion

147

resulting in inferior and supero‐anterior shifts, respectively. The authors concluded

that the multi‐centric nature of functional sinus origin was mediated through a

complex system of separate atrial pacemakers with differential sensitivities to

autonomic control (Boineau et al. 1980). However, while dismissing the presence of

specialised conduction tissue, they acknowledged the potential role of preferential

pathways for the rapid dispersion of the sinus impulse. In humans, Boineau et al.

(1988) reported uni‐focal nodal and extra‐nodal sinus origins in addition to multi‐

centric origins over two to four widely distributed pacemaker sites. The functional

sinus pacemaker complex in humans was estimated to exist over a zone centred about

the long axis of the sulcus terminalis encompassing a 75 by 15mm area, with a cranial

and caudal border of the superior and inferior vena cava, respectively. They suggested

a model where neural and hormonal factors influence both the site of pacemaker

activation and the point of exit from the sinus node, a hypothesis further supported by

studies suggesting relative electrical isolation of the sinus node from surrounding atrial

myocardium with distinct exit points from the node (Bromberg et al. 1995). Consistent

with these observations, in the current study we observed dynamic variation in the

position of atrial activation in the reference group in response to pacing, with early

activation and break‐out points to surrounding myocardium noticeably more caudal

following pacing.

6.4.3 Atrial Remodelling and Sinus Node Function

Sinus node disease is thought to result from disordered impulse generation within the

sinus node or impaired conduction of the impulse to the surrounding atrial tissue

148

(Ferrer 1968). While differentiation between these two potential aetiologies has been

difficult owing to the absence of an appropriate experimental model, Sanders et al.

(2004a) have performed atrial mapping studies in patients with sinus node disease and

found evidence of a diffuse atrial myopathy as evidenced by low atrial voltage, slowed

atrial conduction and prolonged P wave duration. Interestingly, the sinus node

complex was more often uni‐centric and localised to the low crista terminalis with

evidence of regional conduction slowing and scar. While this study showed that

patients with advanced sinus node dysfunction have evidence of structural atrial

remodelling, other studies have shown that sinus node remodelling may occur in

response to short duration atrial pacing or atrial arrhythmias and under these

circumstances may be reversible. Elvan et al. (1996) demonstrated significant

prolongation of CSNRT and the sinus cycle length after two to six weeks of atrial

fibrillation in dogs. In humans, short durations of rapid atrial pacing resulted in similar

sinus node inhibition (Hadian et al. 2002) and CSNRT was significantly longer in

patients cardioverted from chronic lone atrial fibrillation compared to controls

(Kumagai et al. 1991). Manios et al. (2001) also reported that sinus node inhibition did

not revert within 24 hours after cardioversion. However, Hocini et al. (2003)

demonstrated reversibility of sinus impairment in patients in whom atrial fibrillation

was cured by pulmonary vein isolation when follow‐up was continued out to 6 months.

Remodelling of the sinus node has also been reported in clinical conditions such as

asynchronous ventricular pacing (Sparks et al. 1999), atrial septal defects (Morton et

al. 2003) or congestive heart failure (Sanders et al. 2004b). For the latter, Sanders et al.

reported a more caudal localisation of sinus node complex and greater structural

remodelling at the crista terminalis in heart failure patients when compared to

149

controls, raising the possibility that the sinus node remodelling may be a consequence

of these changes via disruption of preferential pathways or loss of tissue capable of

automaticity.

In the current study we have observed significant impairment of sinus node function

following a median 8 months of continuous atrial flutter, consistent with previous

reports. The findings were of similar type, with impairment of sinus node function

associated with structural change, slowed conduction and caudal displacement of the

sinus node complex, albeit to a lesser extent than that described in sinus node disease

(Sanders et al. 2004a). In addition, these findings extend previous observations by

demonstrating caudal shift of both sites of EA and SBO with longer conduction time

within regions of preferential conduction in patients with atrial remodelling due to

persistent arrhythmia.

6.4.4 Preferential Pathways of Conduction

Preferential conduction exists in the atria due to rapid conduction parallel to the

orientation of non‐specialised myocardial cell fibre bundles (Anderson et al. 1981b;

Spach and Dolber 1986; Saffitz et al. 1994). Higa et al. (2004) used the EnSite array to

demonstrate that in some patients with atrial tachycardia the traditionally mapped

origin – the site from which centrifugal activation occurs to the remaining atria – was

indeed the break‐out point. Detailed inspection of the unipolar electrograms

demonstrated a site of earliest activation (“QS” morphology) with a preferential

pathway of conduction to the break‐out point. These investigators were able to

confirm the existence of the origin with a preferential pathway by curative ablation of

150

the tachycardia at EA and/or along the proximal portion of the path to the break‐out

point. It is tempting to draw parallels between the preferential pathways found in

atrial tachycardia, 57% of which were found at the crista terminalis, and those we have

demonstrated in the sinus node complex. With no histological evidence of specialised

atrial conduction tissue, the mechanisms of preferential pathways in atrial tachycardia

and in the sinus node complex is therefore likely to be “functional” relying on weak

cell‐to‐cell coupling to allow conduction of intrinsic activity in an environment

electrically isolated from the surrounding myocardium. This would provide an essential

degree of redundancy to the pacemaker complex, so that while spontaneous

depolarisation at a physiologically appropriate rate continues, several potential exit

sites exist to convey the wave‐front to the myocardium, perhaps with varying

conduction properties determined by the degree of cell‐to‐cell coupling. The

mechanism by which more caudal activation is seen with atrial remodelling, as

demonstrated in this study and others (Sanders et al. 2004b; Sanders et al. 2004a), is

uncertain. We hypothesise that a global remodelling process may affect

proportionately more of a shorter superior preferential pathway than a longer inferior

pathway, leaving superior pathways more susceptible to early development of

conduction abnormalities. This study has demonstrated marked atrial remodelling in

chronic atrial flutter with site‐specific conduction delay and significant voltage

reduction uniformly along the crista terminalis. Although it is feasible that the more

caudal activation is due to impaired automaticity at the superior crista terminalis, this

would require selective remodelling in this region which was not seen in the current

study.

151

6.4.5 Implications

This study demonstrates that the functional sinus node complex is a dynamic and

intricate entity. This complexity is required to balance the need for relative electrical

isolation in order to generate an intrinsic heart beat without the influence of

surrounding myocardium, against the need to conduct such activity to the atrium in a

reliable and consistent fashion. The hypothesis of Boineau et al. (1988) whereby neural

and hormonal factors influence both the site of pacemaker activation and the point of

exit from the sinus node is further supported by our observation of shifting points of

initial activity and exit sites in response to pacing. This study extends the

understanding of the sinus node complex by demonstrating the existence of

preferential pathways of conduction contributing to the diffuse sinus node complex. In

addition, it demonstrates that patients known to have impaired sinus node function

(Sparks et al. 2000; Daoud et al. 2002) have structural remodelling which diminishes

sinus node variability, slows conduction time along preferential pathways, inhibits

conduction across the crista terminalis and results in circuitous conduction of the sinus

impulse.

6.4.6 Limitations

Several studies have now validated the morphology characteristics of virtual

electrograms determined from the EnSite mapping system demonstrating them to

match with high reliability those of conventional contact unipolar electrograms

(Schilling et al. 1998; Thiagalingam et al. 2004b). The technology employed has specific

152

limitations addressed by previous studies and this study has endeavoured to minimise

any inaccuracy by staying within reported limits of accuracy.

6.5 Conclusion

The functional sinus node complex displays a dynamic range of activation along the

postero‐lateral right atrium. Preferential pathways between the site of earliest

activation and conduction to atrial myocardium exist. Remodelled atria demonstrate

more caudal activity, longer conduction time along preferential pathways, diminished

shifts in activation along a constrained functional sinus node complex, lower voltage,

transverse conduction block along the crista terminalis and resultant circuitous wave‐

front propagation. These findings provide additional insight into the function of the

highly complex human sinus node in both health and disease.

153

Table 6.1


Reference (n=15)

Flutter (n=16)

p value

Age (years) 56.9±7.4 60.1±5.2 0.2



Left atrial area (cm2) 18.8±4.5 24.8±5.7 0.01

Right atrial area (cm2) 15.1±2.8 20.2±4.5 0.01


LV end diastolic diameter (mm) 49±6 53±8 0.2

LV = left ventricular

154

Table 6.2

Electrogram characteristics at the crista terminalis catheter under various conditions

Type of beat Reference (n=15)

Flutter (n=16)

p value

Sinus rhythm

2.1±1.8 5.4±2.9 0.002

Pacing at 600ms

1.8±1.5 6.4±3.0 <0.001

Pacing at 450ms

2.1±1.8 6.7±3.1 <0.001

Immediately following 600ms drive train

1.9±1.5 6.1±3.0 <0.001


1.8±1.6 6.4±3.2 <0.001

Number of bipoles demonstrating double potentials or fractionated electrograms on the crista terminalis catheter

With shortest captured extra‐stimulus

4.0±2.6 6.7±3.4 0.03

Sinus rhythm

60.3±14.7 72.4±14.3 0.04

Pacing at 600ms

55.5±10.2 97.2±31.3 <0.001

Pacing at 450ms

58.9±11.4 96.5±30.7 0.001


59.2±13.3 76.2±13.3 0.002


59.7±15.0 80.1±18.5 0.004

Maximum duration of the longest electrogram on the crista terminalis catheter (ms)

With shortest captured extra‐stimulus

91.2±61.7 104±30.8 0.5

155

Table 6.3

Unipolar voltage recorded by the multi‐polar balloon array

Reference (n=15)

Flutter (n=16)

p value

Mean right atrial voltage (mV) 1.78±0.63 1.36±0.63 0.1

Maximum right atrial voltage (mV) 4.96±1.84 3.88±2.07 0.2

Mean voltage at crista (mV) 1.94±0.62 1.28±0.61 0.009

Maximum voltage at crista (mV) 3.50±1.39 2.16±1.20 0.01

Mean voltage at point of EA (mV) 0.12±0.10 0.12±0.13 0.9

Mean voltage at point of SBO (mV) 3.20±1.52 1.54±1.37 0.005

156

Figure 6‐1


Right anterior oblique view of the multi‐polar balloon array in the superior right

atrium, the decapolar catheter in the coronary sinus and the duodecapolar catheter

along the crista terminalis with its stabilising sheath.

157

Figure 6‐2

Method of measuring the position of Earliest Activation and Sinus Break‐Out


158


A right anterior oblique view showing 3 cross‐sectional intra‐cardiac ultrasound views

at the indicated levels of the right atrium. The superior vena cava‐right atrial junction

was defined using ultrasound at the level where the superior lateral crista terminalis is

in proximity to the orifice of the right atrial appendage (middle panel). The crista

terminalis was then marked along its length with vertical line. Vertical measurement

from the superior end of the crista terminalis parallel to this line gives the distance

from the superior vena cava‐right atrial junction. Horizontal measurement

perpendicular to this line gives the anterior distance.

CT – crista terminalis, FO – fossa ovalis, LA – left atrium, RAA – right atrial appendage,

RA – right atrium, SVC – superior vena cava.

159

Figure 6‐3

Example preferential pathway of conduction

Composite map of wave‐front activation from earliest activation (1) via a preferential

pathway (2,3) to sinus break‐out (4) then diverging to the remaining atrium in a patient

with flutter. Electrograms demonstrate a QS pattern at (1) and an RS pattern at (4).

160

Figure 6‐4

Electrograms from the crista terminalis

Example electrograms from a reference patient (left) and a patient with flutter (right)

demonstrating the extent and duration of electrograms with fractionated signals or

double potentials.

II – Surface ECG lead II. CT – crista terminalis. CS – coronary sinus.

161

Figure 6‐5

Distance from superior vena cava ‐ right atrial junction to Earliest Activation and

Sinus Break‐Out points at baseline and following pacing

For the reference group, both Earliest Activation and Sinus Break‐Out points are

relatively superior during sinus rhythm. Immediately following pace‐suppression a

caudal shift is noted in both parameters. For the remodelled group with atrial flutter,

Earliest Activation and Sinus Break‐Out points are located more caudally during sinus

rhythm and the caudal shift is diminished. *p<0.05 for comparisons indicated.

162

Figure 6‐6

Example right atrial maps of Earliest Activation and Sinus Break‐Out

Left map shows earliest activation (1) and sinus break‐out (4) during sinus rhythm with

accompanying electrograms demonstrating a QS morphology at electrogram 1 and an

RS morphology at 4. Right map shows the same for the beat immediately following a

30‐second pacing train in the same patient. Note the caudal displacement of earliest

activation and sinus break‐out.

163

Figure 6‐7

Right atrial circuitous depolarisation pattern


II

1

2

3

4

5

II

1

2

3

4

5

1 2

3 4

164


Typical activation sequence of the circuitous depolarisation pattern seen in patients

without transverse conduction through the crista terminalis. Following break‐out from

the sinus node complex, activation proceeds down the anterior border of the crista

terminalis and then down the posterior border due to the conduction barrier.

1 (top left) Earliest activation (QS) at virtual electrode 1

2 (top right) Sinus break‐out (RS) at virtual electrode 4

3 (bottom left) Bi‐directional depolarisation both inferiorly and supero‐posteriorly

4 (bottom right) Continuing inferior movement of depolarisation either side of crista

terminalis (note linear double potentials at crista terminalis)

165

Movie 6.1

Example of near‐simultaneous dual Sinus Break‐Out points

Please see files on CD supplied entitled Movie 6.1. This movie has been provided in

.mpg, .avi and .wmv formats. The latter is the original version and plays with highest

quality.

The initiation of 3 consecutive beats is shown. Virtual electrograms 1 to 4 placed along

the preferential pathway from EA are displayed below the geometry. The “tracking

virtual” feature of the EnSite tracks the instantaneous peak negative potential along

this path for illustrative purposes. Break‐out to the atrial myocardium is seen at

position 4 and, within 5ms, at position 1 on the crista terminalis near‐simultaneously.

166

Chapter 7. High‐density mapping of atrial fibrillation to determine the optimal recording duration for dominant frequency and

automated detection of complex fractionated electrograms

7.1 Introduction

The understanding of the pathophysiology of atrial fibrillation has evolved such that

ablation targeting the pulmonary veins is highly effective in treating many patients

with paroxysmal atrial fibrillation. However, some patients with paroxysmal and most

patients with persistent atrial fibrillation still require considerable additional substrate

modification to improve ablation outcomes. Areas of CFAE and high DF have been

proposed as critical regions maintaining atrial fibrillation (Berenfeld et al. 2000;

Nademanee et al. 2004; Sanders et al. 2005a; Lin et al. 2006). CFAE have been

implicated as identifying areas acting as pivot points, slowed conduction or anisotropy

all capable of sustaining reentry (Cosio et al. 1986; Spach and Dolber 1986; Konings et

al. 1994). Similarly, localised regions of high frequency activity demonstrating spatio‐

temporal periodicity have been suggested to act as drivers of atrial fibrillation

(Karagueuzian et al. 1998; Skanes et al. 1998; Jalife et al. 2002; Everett et al. 2006).

Thus, identifying such electrograms may allow more targeted ablation to improve

outcomes. To date the evaluation of CFAE has been subjective and development of

automated analysis is paramount for reproducible and objective assessment.

Algorithms for quantification of electrogram fractionation and real‐time DF calculation

have become available and it is desirable that these process a sufficient density of

points without unduly delaying the ablation procedure. The ideal duration of

electrograms required to reproducibly identify sites of CFAE or highest DF is not

167

known. Prior studies have used lengths of electrogram recordings between 2 and 10

seconds to determine these measures (Mansour et al. 2001; Wu et al. 2002; Lazar et al.

2004; Sahadevan et al. 2004; Ravelli et al. 2005; Sanders et al. 2005a; Lin et al. 2006;

Narayan et al. 2006; Ng et al. 2006; Rostock et al. 2006; Ryu et al. 2006; Sanders et al.

2006; Schuessler et al. 2006; Scherr et al. 2007). While longer electrogram recordings

extend the data collection process unnecessarily, shorter durations may provide an

inaccurate representation of the true degree of fractionation or frequency. This study

aimed to determine the minimum duration of electrogram recording that provides

accurate data without unduly lengthening the data collection process, by exploring the

degree of correlation between progressively shorter electrogram recordings with the

longest exportable data sample length.

7.2 Methods


The study comprised 14 patients with symptomatic drug‐refractory atrial fibrillation

undergoing catheter ablation. All anti‐arrhythmic drugs, with the exception of

amiodarone, were ceased ≥5 half‐lives before the study. Prior to the procedure all

patients received anticoagulation with warfarin (International Normalised Ratio 2‐4)

for ≥6 weeks and underwent trans‐oesophageal echocardiography to exclude left atrial

thrombus. The study protocol was approved by the institutional Clinical Research and

Ethics Committee and all patients provided written informed consent for the

procedure.

168

7.2.2 Electrophysiological Study


utilising midazolam and fentanyl. The following catheters were utilised for mapping: (i)

10 pole catheter (2‐5‐2mm inter‐electrode spacing, Daig Electrophysiology) positioned

within the coronary sinus with the proximal bipole at the coronary sinus ostium as

determined in the best septal left anterior oblique position; (ii) 5‐spline, 20 pole

catheter (1mm electrodes separated by 4‐4‐4 inter‐electrode spacing; PentaRay;

Biosense‐Webster). The PentaRay catheter was stabilised using a long vascular sheath

(Preface, Biosense‐Webster or SL0 Braided, St Jude Medical); and (iii) 3.5mm tip

externally irrigated ablation catheter (Celsius Thermocool; Biosense‐Webster). The left

atrium was accessed using a single transeptal puncture. Following transeptal access

bolus unfractionated heparin was utilised to maintain the activated clotting time

between 250‐350 seconds. Following the mapping, patients underwent tailored, step‐

wise ablation comprising pulmonary vein isolation ± linear ablation lines ± CFAE site

ablation.

7.2.3 High‐Density Bi‐atrial Mapping

Mapping was performed during spontaneous (n=8) or induced (n=6) atrial fibrillation

of ≥10 minutes duration. Three‐dimensional reconstructions of the right and left atria

were made using the NavX (Version 6.0R, St Jude Medical) navigation system. High

density mapping of both atria was performed using the PentaRay catheter by

sequentially acquiring 15 simultaneous bipolar electrograms. After ensuring adequate

endocardial contact by fluoroscopy the catheter was held stationary for 8 seconds; this

169

duration for comparator electrogram was chosen as this is the maximum duration

electrogram exportable from the system used and is towards the high end of the range

used in previous studies. These 8 second electrograms were sampled at 1200Hz, band

pass filtered between 30‐500Hz and their location on the 3‐dimensional map

annotated using the Diagnostic Landmark Mapping feature of the system. This ensured

all areas of the atria, venae cavae and pulmonary veins were visited at least once and

the spread of points was equal throughout both chambers. All electrograms were

manually reviewed offline with those demonstrating an unacceptable signal to noise

ratio excluded.


CFAE were defined as fractionated potentials exhibiting multiple deflections from the

isoelectric line and were characterised using the NavX "CFE‐mean" contact mapping

tool. CFE‐mean is defined as the average time duration between consecutive

deflections during a specific time period set by the user and recorded at each site.

Deflections are identified by a detection algorithm and annotated on the electrogram

with yellow "tick‐marks" (Figure 7‐1). Detection is based on three criteria, set by the

user, in which the deflection must: 1) exceed an adaptive "Peak‐to‐Peak Sensitivity"

threshold that is set at a reference‐amplitude slightly greater than the magnitude of

baseline noise; 2) possess a "downstroke" morphology in which the leading local‐

maximum and the trailing local‐minimum amplitude occurs within a time "Duration"

that is set to avoid detection of broad, far‐field events; and 3) exceed a "Refractory"

period from the previous detection that is set to avoid multiple detections on a single

170

deflection. When all three of these criteria are met, a yellow tick‐mark is placed on the

electrogram at the instant of maximum‐negative slope on the downstroke (Figure 7‐1).

In our study, CFE‐mean values were calculated from the maximum possible recording

duration of 8 seconds and exported for analysis by an external computer program.

Also, 7, 6, 5, 4, 3, 2 and 1 second recordings were sub‐sampled and exported for

comparison with the original 8 second recordings. Our user settings included a

Refractory Period of 30ms, a Peak‐to‐Peak Sensitivity typically between 0.05 and 0.1

mV and a Duration of 10ms. Sites with a CFE‐mean value within the range of 40 and

250ms were included for analysis. This wide range was selected to determine

comparative accuracy across the entire spectrum of potentially physiologically relevant

values at differing electrogram durations.


All electrograms were exported at 1200Hz to perform spectral analysis using offline

custom designed software. The electrograms were initially rectified, filtered using

Butterworth filters (Low pass 20Hz, High pass 1Hz) and edge‐tapered with a Hanning

window. Frequency spectra were calculated by Fast Fourier Transform and zero‐

padded for shorter electrogram durations, giving a consistent spectral resolution of

0.07Hz. DF was defined as the frequency containing maximum power within the

frequency domain 3‐15Hz (Figure 7‐2). Points with a Regularity Index (defined as the

ratio of the power within a 0.75Hz band at the DF to total power over 3‐15Hz) of less

than 0.2 calculated from 8 second electrograms were excluded from analysis (Sanders

171

et al. 2005a). DF values were calculated from the entire 8 second sample and from 7,

6, 5, 4, 3, 2 and 1 second sub‐samples for comparison.


Variables are reported as mean ± standard deviation. Left and right atrial CFE‐mean

data were compared using an independent samples t test. Comparison of CFE‐mean

data from 7, 6, 5, 4, 3, 2, and 1 second electrograms with CFE‐mean data from the

index 8 second electrograms from the same points were compared using analysis of

variance with Scheffe’s post hoc comparison. Correlation of CFE‐mean values from 7,

6, 5, 4, 3, 2 and 1 second electrograms with the CFE‐mean values of the index 8 second

electrograms from the same points were quantified using Intraclass Correlation

Coefficients (ICC). Left and right atrial DF comparisons and correlation of sub‐sampled

values with index electrogram values were compared in the same fashion. Statistical

tests were performed using SPSS 13.0 (SPSS Inc).

7.3 Results

7.3.1 Baseline details

Baseline patient characteristics, echocardiographic data and medications are

summarised in Table 7.1. Electrograms from a total of 6967 points (498±174

points/patient) were analysed; 228±102 points/patient in the right and 270±86

points/patient in the left atria. Bi‐atrial mapping took 37±13 minutes to complete.

172

7.3.2 CFE‐mean

For the index 8 second samples CFE‐mean was 114±20ms for the right atria and

102±17ms for the left atria (p=0.01). Bi‐atrial mean values for all electrogram durations

are shown in Figure 7‐3A. A trend to decrease with shorter electrograms is seen but

these are not significantly different to the index mean for all points of 109ms (p=0.2).

However, it is point‐by‐point accuracy that is more important for CFE‐mean and under‐

estimations tend to cancel out over‐estimations in mean values. The mean absolute

differences between 8 second electrogram index CFE‐mean and the 7, 6, 5, 4, 3, 2 and

1 second CFE‐mean are therefore shown in Figure 7‐3B. The increase in the magnitude

of absolute difference as electrogram duration decreases reflects the poorer

correlation of each shorter sub‐sampled electrogram value with its index 8 second

value. The percentage of sub‐sampled electrograms where CFE‐mean deviated from

the 8 second index CFE‐mean by more than 10% is shown in Figure 7‐4. Progressively

more points demonstrate >10% difference from the index value as electrogram

duration shortens. Scatter plots of CFE‐mean for the index 8 second value (y axis)

versus the CFE‐mean for the same point derived from the 7, 6, 5, 4, 3, 2 and 1 second

sub‐samples (x axes) are shown in Figure 7‐5A. Intraclass Correlation Coefficients of

sub‐sampled electrogram CFE‐mean as compared to the index 8 second electrogram

CFE‐mean quantify this weakening relationship as the electrogram duration becomes

shorter.

173


For the index 8 second samples DF was 5.7±0.8Hz for the right atria and 6.0±0.8Hz for

the left atria (p=0.02). Bi‐atrial mean values for all electrogram durations are shown in

Figure 7‐3A. Means for 4 second samples or shorter are significantly different to the 8

second index mean for all points of 5.9Hz (p<0.001). The mean absolute differences

between 8 second electrogram index DF and the 7, 6, 5, 4, 3, 2 and 1 second DF are

shown in Figure 7‐3B. Similarly, the increase in mean absolute difference as

electrogram duration decreases reflects the poorer correlation of each shorter sub‐

sampled electrogram value with its index 8 second value. The percentage of sub‐

sampled electrograms where DF deviated from the 8 second index DF by more than

10% is shown in Figure 7‐4. Progressively more points demonstrate >10% difference

from the index value as electrogram duration shortens. Scatter plots of DF for the

index 8 second value (y axis) versus the DF for the same point derived from the 7, 6, 5,

4, 3, 2 and 1 second sub‐samples (x axes) are shown in Figure 7‐5B. Intraclass

Correlation Coefficients of sub‐sampled electrogram DF as compared to the index 8

second electrogram DF quantify this weakening relationship as the electrogram

duration becomes shorter.

7.4 Discussion


This study presents new information about the duration of electrogram required to

efficiently and accurately determine the degree of fractionation and DF. Recordings of

174

shorter durations were compared to the maximum duration recordable by the system

using the following criteria: i) mean values, ii) mean absolute differences of >10ms or

0.5Hz, iii) proportion of points deviating >10% from the index value, and iv) ICC. CFE‐

mean values derived from electrogram durations of 3 seconds or less have a mean

absolute error of more than 10ms (Fig 3B) and for electrogram durations of 4 seconds

or less, at least 1 in 5 points deviate more than 10% from the index value (Fig 4).

Intraclass Correlation Coefficients for CFE‐mean demonstrate “substantial agreement”

(ICC>0.8) for all sub‐samples except 1 second electrograms. Visual inspection of CFE‐

mean scatter plots suggests electrograms of more than 4 seconds approximate the

index 8 second electrogram value (Fig 5A). Mean DF derived from electrograms of 4

seconds or less differ significantly from the index 8 second mean (Fig 3A) and DF values

from electrogram durations of 4 seconds or less have a mean absolute error of more

than 0.5Hz (Fig 3B). With electrogram durations of 3 seconds or less, for at least 1 in 5

points DF deviates more than 10% from the index value (Fig 4). Intraclass Correlation

Coefficients for DF demonstrate “substantial agreement” for 7 second sub‐samples,

“moderate agreement” (ICC 0.6‐0.8) for 6 and 5 second sub‐samples, and only fair or

poor agreement for shorter electrograms. Visual inspection of DF scatter plots

suggests electrograms of more than 4 seconds approximate the index 8 second

electrogram value (Fig 5B).

These observations, based on the above criteria, suggest a minimum of 5 seconds is

required to achieve accuracy comparable to the maximum recordable sample of 8

seconds.

175

7.4.2 Time Domain Analysis

The postulated mechanisms that bring about CFAE include anisotropic propagation

(Spach and Dolber 1986), areas of slowed conduction (Cosio et al. 1983) and multiple

wavelets leading to random or leading circle reentry. CFAE are observed at pivot points

where wavelets rotate at zones of functional block (Konings et al. 1994). Selective

elimination of CFAE by catheter ablation has been proposed to halt wavelet reentry,

thereby preventing perpetuation of atrial fibrillation (Nademanee et al. 2004; Oral et

al. 2007). Previous reports of CFAE mapping have used descriptive terms to

subjectively determine which electrograms fit these criteria. Nademanee et al. (2004)

defined CFAE as fractionated electrograms composed of ≥2 deflections, perturbation

of the baseline with continuous deflection of a prolonged activation complex, or atrial

electrograms with a cycle length ≤120ms. Rostock et al. (2006) used fractionated

potentials with ≥3 deflections from the isoelectric line or continuous activity as their

criteria for CFAE. Oral et al. (2007) targeted electrograms with a cycle length ≤120ms or

shorter than the coronary sinus; or those that were fractionated or displayed

continuous electrical activity. Few studies have tried to automate this procedure and

the minimum electrogram duration required to accurately determine electrogram

complexity is unknown. Ravelli et al. (2005) studied fibrillatory wave complexity using

an automated wave similarity algorithm calculated over 5 second windows. Narayan et

al. (2006) used a “gold‐standard” of manually measured atrial fibrillation cycle length

to validate an autocorrelation method of determining local cycle length. Atrial

fibrillation cycle length estimates were more stable performed over 10 seconds rather

than 2 seconds due to short term cycle length fluctuations. Scherr et al. (2007) have

recently used another automated algorithm of 2.5 second recordings to classify 86% of

176

1904 left atrial sites from 19 patients as exhibiting CFAE, one third of which were

deemed “highly repetitive”. No comparison between different electrogram durations

was made.

7.4.3 Frequency Domain Analysis

Regions of high frequency activations demonstrating spatio‐temporal periodicity have

been suggested as local drivers of atrial fibrillation. Such “rotors” or “focal sources”

would be identifiable by fast activity and may be constantly present. These drivers

have been suggested to be present where structural remodelling of the atria has

occurred, and absent where atrial fibrillation is due to electrical remodelling alone

where multiple wavelets predominate (Everett et al. 2006). Further evidence for atrial

fibrillation spatial organisation comes from optical mapping in animals using DF to

demonstrate areas of periodic high frequency (Skanes et al. 1998; Kalifa et al. 2006).

Retrospective analysis of ablation of paroxysmal atrial fibrillation shows that cycle

length prolongation and eventual termination occurs at sites of high DF (Sanders et al.

2005a). Prior studies to localise DF sites have used electrogram sample lengths of

between 2 to 10 seconds (Mansour et al. 2001; Wu et al. 2002; Lazar et al. 2004;

Sanders et al. 2005a; Lin et al. 2006; Narayan et al. 2006; Ng et al. 2006; Ryu et al.

2006; Sanders et al. 2006; Schuessler et al. 2006). Ng et al. (2007) reported good

correlation for DF calculated from 2 second sub‐samples with atrial activation rates

calculated in the time domain from 4 second recordings. Correlation between 1 and 4

second samples was less satisfactory, therefore a 2 second minimum electrogram was

recommended. Additionally, the mean DF of 4 sequential 4 second electrograms

177

showed improved correlation with activation rates compared to the DF of a single 4

second electrogram, both for simulated and clinical electrograms (Ng et al. 2006).

Harmonics in the frequency spectrum are a source of error in the estimation of true

atrial activation rate in the highly fractionated signal (Ng et al. 2006; Ryu et al. 2006).

The use of indices such as the Regularity Index used here has sought to eliminate such

conflicting information from analysis (Sanders et al. 2005a).

7.4.4 Differences in Effect of Electrogram Duration between CFE‐mean and DF

For DF, shorter electrogram durations tend to over‐estimate the value, predominantly

due to more prominent harmonics in the frequency spectra. This can be seen in Figure

7‐5B where an additional “line” at half the slope to the right of the regression line

becomes more prominent as electrograms shorten. This explains the poorer ICC values

for DF than CFE‐mean; a significant number of points have twice the index value,

whereas CFE‐mean errors are more homogenously distributed around the regression

line (Figure 7‐5A). The harmonics also explain why shorter electrograms provide a

significantly higher mean DF, whereas the average CFE‐mean does not significantly

change with shorter electrogram durations. Despite better ICC, CFE‐mean has more

points deviating >10% from the index value for electrograms of 5 seconds or less,

compared with DF (Figure 7‐4). This suggests that although errors in DF are less

common from shorter electrograms, they are likely to be greater in magnitude.

178

7.4.5 Limitations

The results of this study pertain to the specific algorithms and pre‐processing methods

employed. CFAE‐guided ablation has been established as an effective treatment for

atrial fibrillation, but the definitions of targets differ. The CARTO system (Biosense‐

Webster) has recently added an automated fractionation detection algorithm to allow

real‐time annotation of fractionated sites (CFAE Software Module) using a fixed 2.5

second interval (Scherr et al. 2007). This algorithm differs to that employed in this

study, in that it includes both low and high voltage thresholds (default 0.05 and

0.15mV), alternative interval thresholds (default 70 to 120ms), and indicates the

degree of fractionation by Interval Confidence Level (intervals detected per 2.5

seconds). Given the differences between this algorithm and the one employed in this

study, extrapolation of this study’s results regarding minimum electrogram duration to

other CFAE detection algorithms may not be accurate. Finally, while this study aimed

to determine the ideal recording duration for CFAE and DF analysis, it was not

designed to evaluate the clinical efficacy of these parameters (or software) to guide

ablation.

7.5 Conclusion

The minimum electrogram duration for accurate analysis of CFAE and DF by the

methods described is 5 seconds. This is based on the comparison of values derived

from electrograms of 1 to 7 seconds and their comparison with 8 second electrogram

values. Electrograms of 4 seconds or less have unacceptable rates of error by the

179

methods used and may be misleading in the ablation of atrial fibrillation guided by

these parameters.

180

Table 7.1


Number of Patients 14

Age (years) 57±11

History of atrial fibrillation (months) 68±50

Type of atrial fibrillation: Paroxysmal/Persistent 5/9

Structural heart disease

Ischaemic heart disease 1

Hypertrophic cardiomyopathy 1

Hypertension 3

Left atrial dimension (mm) 48±6

Left atrial area (cm2) 29±6

Right atrial area (cm2) 24±7

Inter‐ventricular septal thickness (mm) 11±2

Left ventricular end diastolic dimension (mm) 52±5

Left ventricular ejection fraction (%) 57±9

Anti‐arrhythmic preceding electrophysiology study

Amiodarone 4

Sotalol 5

Flecainide 2

181

Figure 7‐1

Composite image of bi‐atrial high‐density CFE‐mean colour map, surface ECG and

bipolar atrial electrograms


182


A short example electrogram (centre) is given demonstrating 2 yellow tick‐marks on

downstrokes between leading local‐maximum and trailing local‐minimum points. Tick‐

marks along a full 8 second electrogram (bottom) indicate annotations of deflection

events in the electrogram according to the Refractory, P‐P sensitivity and Duration

criteria defined in the text. CFE‐mean is the average time duration between

annotations and the value for each point is displayed along a colour spectrum on the 3‐

dimensional geometry (top).

AP – antero‐posterior, PA – postero‐anterior, SVC – superior vena cava, LSPV – left

superior pulmonary vain, RSPV – right superior pulmonary vein, LAA – left atrial

appendage, LIPV – left inferior pulmonary vein, RIPV – right inferior pulmonary vein,

IVC – inferior vena cava, LA – left atrium, RA – right atrium.

183

Figure 7‐2

Example electrogram with accompanying frequency spectra

Frequency spectra from 3 to 15Hz derived by Fast Fourier Transform analysis from 8, 7,

6, 5, 4, 3, 2 and 1 second samples of the 8 second bipolar electrogram shown at top.

Dominant frequency is the frequency with the highest power (y axis).

DF – dominant frequency, RI – regularity index.

184

Figure 7‐3

Effect of electrogram duration on CFE‐mean and DF

(A)

5.8

6.0

6.2

6.4

6.6

8s 7s 6s 5s 4s 3s 2s 1s106

107

108

109

110

111

DFCFE-mean

Dom

inan

t Fre

quen

cy (H

z)

Time (s)

CFE-m

ean (ms)

**

*

*

(B)

0.00

0.25

0.50

0.75

1.00

1.25

7s 6s 5s 4s 3s 2s 1s0

5

10

15

20

25

DFCFE-mean

Dom

inan

t Fre

quen

cy (H

z)

Time (s)

CFE-m

ean (ms)

Effect of electrogram duration for CFE‐mean and DF (mean ± standard error of the

mean) for (A) mean values and (B) mean absolute differences between sub‐sampled

electrogram values and index 8 second values.

*p<0.001 for comparison with index 8 second mean.

185

Figure 7‐4

Effect of electrogram duration on accuracy

0

10

20

30

40

50

60

7s 6s 5s 4s 3s 2s 1s

CFE-mean Dominant Frequency

%

Time (s)

Percentage of electrograms deviating by more than 10% from the index 8 second value

for CFE‐mean and DF. Both parameters show a greater proportion of inaccurate values

as electrogram duration shortens.

186

Figure 7‐5

Scatter plots of CFE‐mean and DF

(A) (B)

Scatter plots of Inde

x 8 second

value

s (y axis) versus the values fo

r the same po

int d

erived

from

the 7, 6, 5, 4, 3, 2

and

1 second

sub

‐sam

ples (x axes) for (A) CFE‐mean and (B) DF. Note the de

creasing

Intraclass Correlatio

n Co

efficients

(ICC) with

sho

rter electrogram

durations. A

value

of 1

indicates pe

rfect a

greemen

t and

0 indicates no

agreemen

t.

187


relationship between activation frequency, complex fractionated electrograms and the anatomical substrate

8.1 Introduction

There have been significant advances in our understanding of the pathophysiology of

atrial fibrillation over the last decade. These have led to ablation targeting the

pulmonary veins being able to abolish atrial fibrillation in a significant proportion of

patients with paroxysmal atrial fibrillation. However, some patients with paroxysmal

and most patients with persistent atrial fibrillation still require considerable additional

substrate modification to improve outcomes. Areas of CFAE and high DF have been

proposed as critical regions maintaining atrial fibrillation (Berenfeld et al. 2000;

Nademanee et al. 2004; Sanders et al. 2005a). The postulated mechanisms that bring

about CFAE include anisotropic propagation, areas of slowed conduction, multiple

wavelets leading to random or leading circle reentry, pivot points where wavelets

rotate about sites of functional block and out of phase activation (Cosio et al. 1983;

Spach and Dolber 1986; Konings et al. 1994). Selective elimination of CFAE by catheter

ablation has been proposed to halt wavelet reentry, thereby preventing the

perpetuation of atrial fibrillation (Nademanee et al. 2004; Oral et al. 2007; Schmitt et

al. 2007; Nademanee et al. 2008). Similarly, localised regions of high frequency activity

demonstrating spatio‐temporal periodicity have been suggested to act as drivers of

atrial fibrillation (Skanes et al. 1998; Jalife et al. 2002; Everett et al. 2006; Lin et al.

2006). Thus, identifying electrograms with these features may allow targeted ablation

to improve success rates. However, there is a paucity of data on the relationship

188

between CFAE and DF. In particular it is not known whether they are markers of the

same or provide complementary information on the pathogenesis of atrial fibrillation.

Understanding the relationship between these targets may be useful in guiding

radiofrequency ablation. In this clinical study we aimed to examine the relationship

between sites of high frequency activation and fractionation during atrial fibrillation.

8.2 Methods


The study comprised 20 patients with symptomatic drug‐refractory atrial fibrillation

undergoing catheter ablation. All anti‐arrhythmic drugs, with the exception of

amiodarone, were ceased ≥5 half‐lives before the study. Prior to the procedure, all

patients received anticoagulation with warfarin (International Normalised Ratio 2‐4)

for ≥6 weeks and underwent trans‐oesophageal echocardiography to exclude left atrial

thrombus. The study protocol was approved by the institutional Clinical Research and

Ethics Committee and all patients provided written informed consent for the

procedure.

For the purposes of the study atrial fibrillation was defined in accordance with the HRS

Expert Consensus Statement (Calkins et al. 2007); paroxysmal atrial fibrillation was

defined as atrial fibrillation episodes that terminate spontaneously within 7 days and

persistent as having greater than 7 days of atrial fibrillation. Indeed, all patients with

persistent atrial fibrillation in this study had a minimum of 6 months continuous atrial

fibrillation at some point prior to their study.

189



utilising midazolam and fentanyl. The following catheters were utilised for mapping: (i)

10 pole catheter (2‐5‐2mm inter‐electrode spacing, Daig Electrophysiology) positioned

within the coronary sinus with the proximal bipole at the coronary sinus ostium as

determined in the best septal left anterior oblique position; (ii) 5‐spline, 20 pole

catheter (1mm electrodes separated by 4‐4‐4mm inter‐electrode spacing; PentaRay;

Biosense‐Webster). The PentaRay catheter was stabilised using a long vascular sheath

(Preface, Biosense‐Webster or SL0 Braided, St Jude Medical); and (iii) 3.5mm tip

externally irrigated ablation catheter (Celsius Thermocool, Biosense Webster). The left

atrium was accessed using a single transeptal puncture. Following transeptal access,

bolus unfractionated heparin was utilised to maintain the activated clotting time

between 250‐350 seconds.

After mapping, patients underwent circumferential pulmonary vein ablation with an

end‐point of isolation confirmed by circumferential mapping (Lasso, Biosense‐

Webster) with either elimination or dissociation of pulmonary venous potentials.

Ablation of the pulmonary veins was performed using a delivered power of 30 W with

irrigation rates of 30mL/min. Additional substrate modification by linear ablation (roof‐

line or mitral isthmus ablation) or targeting regions of CFAE was performed in patients

with atrial fibrillation episodes persisting for > 48 hours, structural heart disease or

with marked left atrial dilatation (longest diameter >57mm). Cavotricuspid isthmus

ablation with an endpoint of bi‐directional isthmus block was performed only in

patients with a history of typical atrial flutter or if mapping confirmed typical atrial

190

flutter during the procedure. The endpoint of substrate modification was either

electrophysiologically confirmed linear conduction block established via pacing

manoeuvres or the elimination of local fractionation. Substrate modification was

performed using a delivered power of 30‐35 W with irrigation rates of 30‐60mL/min.

Detailed DF or CFE‐mean analyses were not available to guide the procedure.

8.2.3 Study Protocol

Mapping was performed during spontaneous or induced (n=9) sustained atrial

fibrillation of >10 minutes duration. Three‐dimensional reconstructions of the right

and left atria were made using the NavX navigation system (Version 6.0R or 7.0, St

Jude Medical). High‐density mapping of both atria was performed using the PentaRay

catheter by sequentially acquiring 15 simultaneous bipolar electrograms. After

ensuring adequate endocardial contact by fluoroscopy the catheter was held

stationary for 8 seconds. These 8 second electrograms were sampled at 1200Hz, band

pass filtered between 30‐500Hz and their location on the 3‐dimensional map

annotated using the Diagnostic Landmark Mapping feature of the system. This ensured

all areas of the atria, venae cavae and pulmonary veins were visited at least once and

aimed to have the spread of points equal throughout both chambers. All electrograms

were manually reviewed offline with those demonstrating an unacceptable signal to

noise ratio excluded from analysis. For the purpose of regional evaluation, points were

designated as belonging to one of 4 right atrial areas (anterior, septal, posterior or

lateral) or 8 left atrial areas (roof, anterior, septum, lateral, posterior, inferior,

pulmonary veins or appendage) using previously validated software (Coperniqs,

191

Medicalgorithmics) (Kuklik et al. 2004). Points designated at pulmonary veins were

well within the veins, not at the ostia or antra which were considered part of a left

atrial area.

The baseline atrial fibrillation cycle length within the coronary sinus was determined as

previously described by averaging intervals over 1 minute using automated cycle

length monitoring software (Bard Electrophysiology). Inter‐electrogram intervals of

<100ms and continuous electrical activity were counted as a single interval. At each

time point, the automated annotation was manually verified and corrected with online

calipers at a paper speed of 100mm/s (Sanders et al. 2005a).


CFAE were defined as fractionated potentials exhibiting multiple deflections from the

isoelectric line. For the purposes of the study, CFAE were quantified using the

automated NavX "CFE‐mean" contact mapping tool, recently shown to effectively

guide CFAE‐targeted ablation (Verma et al. 2008), which provided a measure of

electrogram fractionation as the average time duration between consecutive

deflections during the 8 second electrogram. Deflections were identified by a

detection algorithm and annotated on the electrogram with yellow "tick‐marks".

Detection is based on three user‐defined criteria in which the deflection must: 1)

exceed an adaptive "Peak‐to‐Peak Sensitivity" threshold set at a reference‐amplitude

slightly greater than baseline noise; 2) possess a "downstroke" morphology in which

the leading local‐maximum and the trailing local‐minimum amplitude occurs within a

time "Duration" that is set to avoid detection of broad, far‐field events; and 3) exceed

192

a "Refractory" period from the previous detection that is set to avoid multiple

detections on a single deflection. When all three of these criteria are met, a yellow

tick‐mark is placed on the electrogram at the instant of maximum‐negative slope on

the downstroke (Figure 8‐1). Our user settings were a Refractory Period of 30ms, a

Peak‐to‐Peak Sensitivity of 0.1mV and a Duration of 10ms. Points with a CFE‐mean

value from the minimum of 40ms (Refractory + Duration) up to 250ms were included

for analysis to ensure inclusion of all potentially relevant data.

8.2.5 Activation Frequency

All electrograms were exported at 1200Hz to perform spectral analysis using offline

software (Coperniqs). The electrograms were initially rectified, filtered using

Butterworth filters (Low pass 20Hz, High pass 1Hz) and edge‐tapered with a Hanning

window. Frequency spectra were calculated by Fast Fourier Transform with a spectral

resolution of 0.07Hz. DF was defined as the frequency containing maximum power

within the frequency domain 3‐15Hz. Points with a Regularity Index (defined as the

ratio of the power within a 0.75Hz band at the DF to total power over 3‐15Hz) of less

than 0.2 were excluded from analysis (Sanders et al. 2005a).

8.2.6 Relationship between Activation Frequency and Fractionation

The relationship between CFAE and DF were evaluated as follows: (i) Point‐by‐point, to

explore correlation at each sample location in every patient; (ii) Patient‐by‐patient, to

explore correlation between medians of CFE‐mean and DF for a bi‐atrial map; and (iii)

193

Spatial distribution, to examine the relative positions of fractionation to high DF points.

The latter was determined by displaying the 20 highest DF points for each patient and

grouping them in clusters (for points <5mm apart) showing a frequency gradient to the

surrounding atria as previously described (Sanders et al. 2005a). The distance from

these clusters to areas of low CFE‐mean (multiple contiguous points within 5mm

demonstrating CFE‐mean<80ms) was measured on the 3‐dimensional maps. In

addition, activation mapping was performed on the 15 simultaneously acquired

PentaRay bipolar electrograms at sites of high DF. Activation mapping was only

performed during periods of organised atrial fibrillation as previously described

(Haïssaguerre et al. 2006b). Local activation was manually annotated to the peak or

trough of the largest amplitude deflection on the bipolar contact electrogram.




percentage. Proportions were compared using Fisher’s Exact test. Comparisons were

analysed using paired or unpaired t tests, or the Mann‐Whitney U test as appropriate.

To analyse differences in CFE‐mean or DF according to atrial fibrillation type and

region, a linear mixed effects model was fitted to the data. In the model, atrial

fibrillation type, region, and the interaction between atrial fibrillation type and region

were fitted as fixed effects. Random effects were included in the model to account for

the dependence in observations both within a patient and within a given region of a

patient, following which post‐hoc pair‐wise comparisons were performed within the

194

model. Correlation between variables was measured by Pearson’s r statistic. Statistical

significance was set at p<0.05.

8.3 Results


There were 10 patients with paroxysmal and 10 with persistent atrial fibrillation.

Baseline patient characteristics are summarised in Table 8.1. Structural heart disease

was present in 2 patients, 1 with ischaemic and the other with hypertrophic

cardiomyopathy. In addition, 5 patients had hypertension, 2 had prior transient

cerebrovascular events and 1 had obstructive sleep apnoea. Left atrial size and left

ventricular hypertrophy were greater in the persistent compared to paroxysmal atrial

fibrillation group.

Electrograms from a total of 10,133 points were analysed (507±150 points/patient);

238±79 points/patient in the right and 268±102 points/patient in the left atria. Atrial

fibrillation cycle length as measured in the coronary sinus at the commencement of

the study was 174±20ms (184±20ms for paroxysmal atrial fibrillation, 163±14ms for

persistent atrial fibrillation, p=0.02).


Median CFE‐mean was 103ms (IQR 76‐143ms) for right atrial points and 98ms (IQR 72‐

139ms) for left atrial points (p<0.001). Median CFE‐mean was 117ms (IQR 85‐161ms) in

195

paroxysmal atrial fibrillation and 92ms (IQR 70‐127ms) in persistent atrial fibrillation

(p<0.001). Figure 8‐2A demonstrates the regional variation in CFE‐mean, with high

CFE‐mean values indicating less fractionation. CFE‐mean was significantly lower in

persistent atrial fibrillation compared to those with paroxysmal atrial fibrillation for all

12 regions (p=0.009). The difference between regions was not dependent on type of

atrial fibrillation by the statistical model employed (p=0.7), hence the comparisons

presented between areas for CFE‐mean are across all patients, regardless of atrial

fibrillation type. The areas of greatest fractionation within the left atrium were the

roof, septum, posterior, anterior and inferior walls. All right atrial regions

demonstrated a higher CFE‐mean than these areas (p<0.05, Figure 8‐2A).

Areas containing multiple points with low (<80ms; lowest quartile) CFE‐mean were

located on the 3‐dimensional maps (Off et al. 2008). The mean number of such areas

per patient was 10.3±4.2 and these areas were compared in location to clusters of high

DF points (see below). More areas of low CFE‐mean per patient were seen in persistent

compared to paroxysmal atrial fibrillation (13±3.5 versus 7.5±3.5; p=0.002).

Median CFE‐mean correlated moderately with atrial fibrillation cycle length (r=0.47,

p=0.04) and there was an inverse correlation (r=‐0.45, p<0.05) between mean atrial

fibrillation cycle length and the number of areas of low CFE‐mean.


Median DF was 5.49Hz (IQR 4.91‐6.15) for right atrial points and 5.57Hz (IQR 4.91‐6.23)

for left atrial points (p=0.02). Median DF was 5.27Hz (IQR 4.69‐5.71) in paroxysmal

atrial fibrillation patients and 5.93Hz (IQR 5.35‐6.67) in persistent atrial fibrillation

196

patients (p<0.001). Regional differences in DF are illustrated in Figure 8‐2B. DF was

higher in persistent atrial fibrillation for all regions compared to paroxysmal atrial

fibrillation (Figure 8‐2B). The differences between regions across type of atrial

fibrillation were not uniform (p=0.01), hence the comparisons presented between

regions are made only within paroxysmal or within persistent atrial fibrillation groups.

For paroxysmal atrial fibrillation the DF of pulmonary veins, left atrial appendage and

roof was significantly higher than in the inferior and lateral left atrial regions (p<0.05,

Figure 8‐2B). For persistent atrial fibrillation no significant differences were seen

between left atrial areas, except that the inferior left atrium had lower DF than the 3

highest areas; lateral and posterior left atrium, and left atrial appendage (p<0.05,

Figure 8‐2B).

The 20 highest DF points for each patient were located on the 3‐dimensional maps.

These 20 highest DF points were clustered such that the mean number of high DF

clusters was 5.2±1.7 per patient. These clusters were compared in location to areas of

low CFE‐mean. Of 104 clusters, 68 (65%) were within the left atria, of which 24 were

located in the pulmonary veins.

Median DF correlated well with atrial fibrillation cycle length (r=‐0.67, p=0.001);

however, there was no significant correlation between atrial fibrillation cycle length

and the number of high DF clusters (p=0.3).

8.3.4 Relationship between CFE‐mean and Dominant Frequency

Correlation of CFE‐mean and DF on a patient‐by‐patient basis shows a moderate

inverse correlation (r=‐0.50, p=0.03; Figure 8‐3), suggesting that for a given patient the

197

degree of fractionation (as indicated by low CFE‐mean) increases with the DF.

However, point‐by‐point correlation between CFE‐mean and DF reveals a poorer

correlation (r=‐0.17, p<0.001; Figure 8‐3), suggesting that these two parameters, while

not the same, may co‐exist within a common area. To further explore the relationship

between these measures, the spatial relationship of the 104 clusters of high DF to

CFAE areas were evaluated. The median distance between DF clusters and CFAE areas

was 5mm (IQR 2‐10). Eighty percent of DF clusters were at or less than 10mm from

fractionation and in a further 10% they were found between 10‐20mm. Figure 8‐4

shows a representative example of the spatial relationship of fractionation to a high DF

site. It shows the electrograms and their frequency spectra from high DF points,

together with the electrograms from nearby low CFE‐mean areas. The high DF site

demonstrates a frequency gradient to the surrounding tissue with adjacent

electrograms demonstrating more fractionation. In this example, the distance from the

high DF site to the most fractionated electrogram shown was 7mm.

To further determine the relationship between sites of high DF and CFAE we

performed activation mapping where possible using simultaneous electrograms from

the PentaRay catheter at sites of high DF. Figure 8‐5 demonstrates activation mapping

at high DF cluster sites and reveals that during organised activity, varying patterns of

wave‐front activation and isochronal crowding are seen adjacent to the high DF site.

Electrogram fractionation is observed to occur in close proximity to high DF sites in

these examples.

198

8.4 Discussion


This study utilised high‐density endocardial mapping of atrial fibrillation to describe

the relationship of activation frequency and electrogram fractionation within the atria.


(i) The left atrium is more highly fractionated than the right atrium and patients

with persistent atrial fibrillation demonstrate a higher degree of fractionation

than those with paroxysmal atrial fibrillation.

(ii) Activation frequency is higher in the left atrium than the right atrium and

higher in persistent atrial fibrillation than paroxysmal atrial fibrillation. In

paroxysmal atrial fibrillation regional analysis demonstrated significant

variation in frequency, but these regional differences were less apparent in

persistent atrial fibrillation.

(iii) Fractionation and activation frequency correlate with atrial fibrillation cycle

length. While these variables do not demonstrate a relationship on a point‐by‐

point basis, within an individual a moderate relationship is observed between

them.

(iv) Exploration of their spatial relationship suggests that electrogram fractionation

is observed in close proximity to sites of high frequency activation and is

further supported by high‐density simultaneous activation mapping at these

sites.

199

8.4.2 Spatial Distribution of Complex Fractionated Atrial Electrograms

The finding of a higher degree of fractionation in the left atrium than the right is

consistent with previous reports. Jaïs et al. (1996) initially described regional bi‐atrial

disparities in endocardial fractionation in paroxysmal atrial fibrillation, at a time when

much of the data were from epicardial studies. They found that complex electrical

activity was more frequently found in the septal and posterior regions of both atria,

with more regular activity seen in the trabeculated atria. This group has further gone

on to describe the characteristics of CFAE predictive of favourable ablation targets

(Takahashi et al. 2008). Proponents of CFAE‐targeted ablation have found fractionation

predominantly in left atrial sites: the pulmonary vein ostia; inter‐atrial septum; mitral

annulus; basal left atrial appendage; left atrial roof; and the posterior left atrium

(Nademanee et al. 2004; Oral et al. 2007; Schmitt et al. 2007; Porter et al. 2008). They

reported that such sites were found more widely spread amongst those with longer

duration atrial fibrillation, consistent with the greater number of electrogram

fractionation areas found in the patients with persistent atrial fibrillation in the current

study.

8.4.3 Spatial Distribution of Dominant Frequency

The initial evidence for organisation of atrial fibrillation came from optical mapping in

animal studies which used spectral analysis to demonstrate repetitive high frequency

activity with temporal and spatial periodicity during atrial fibrillation, consistent with

the hypothesis of “driving” rotors (Skanes et al. 1998; Kalifa et al. 2006). More

recently, similar findings have been reported in humans (Wu et al. 2002; Lazar et al.

200

2004; Sahadevan et al. 2004; Sanders et al. 2005a; Haïssaguerre et al. 2006b). Our

finding of higher DF in the left atrium in humans is consistent with previous published

reports on paroxysmal atrial fibrillation (Lazar et al. 2004; Lin et al. 2006).

Furthermore, the maximal DF has been reported in the pulmonary veins in patients

with paroxysmal atrial fibrillation (Sanders et al. 2006). However, persistent atrial

fibrillation had a less conclusive regional pattern of DF distribution. Although previous

publications have shown a higher mean DF in the left atria, particularly around the

pulmonary veins (Wu et al. 2002), others have shown that this distinction disappears

for longstanding atrial fibrillation where other sites are observed to predominate

(Lazar et al. 2004; Atienza et al. 2006). Isolation of the pulmonary veins abolishes this

left‐to‐right gradient in paroxysmal atrial fibrillation and for the subset of patients with

persistent atrial fibrillation in whom a left‐to‐right gradient exists, success of

pulmonary vein isolation alone exceeds that of persistent atrial fibrillation without

such gradient (Lazar et al. 2006). Retrospective analysis of ablation of paroxysmal atrial

fibrillation shows that atrial fibrillation cycle length prolongation and eventual

termination occurs at sites of high DF (Sanders et al. 2005a). This has not been shown

for persistent atrial fibrillation and may be in part due to our study’s finding that DF is

distributed more evenly throughout atrial regions, making targeting any one region a

less effective strategy for ablation in this sub‐group. Indeed, a more aggressive, multi‐

region approach for ablating longstanding persistent atrial fibrillation has been

advocated (Haïssaguerre et al. 2005c).

201

8.4.4 Relationship between Activation Frequency and Fractionation

Few studies have examined the relationship between CFAE and DF. Lemola et al.

(2006) examined the DF before and after atrial fibrillation ablation by 2 techniques;

circumferential‐pulmonary‐vein and electrogram‐guided ablation. Using CFAE to guide

ablation, the mean left atrial DF was seen to reduce by 17% and this decrease was

associated with less recurrent atrial fibrillation for patients with persistent atrial

fibrillation. No such prognostic advantage of DF reduction was seen for the pulmonary

vein encircled group, suggesting that differing mechanisms are eliminated by the 2

ablation strategies. Rostock et al. (2006) have shown that dynamic changes in CFAE are

dependent on regional atrial fibrillation cycle length, in that shortening of local cycle

length preceded development of CFAE. Our finding of significant correlation between

atrial fibrillation cycle length and number of areas of low CFE‐mean, and the proximity

of DF clusters to CFAE areas, is consistent with this observation. A recent study

evaluating the potential role of ganglionated plexi in atrial fibrillation has

demonstrated that following local administration of acetyl choline there was a

significant gradient of decreasing DF and incidence of CFAE from ganglionated plexi

toward distant sites (Lu et al. 2008). While we did not evaluate the role of

ganglionated plexi, this inter‐relationship of CFAE and DF is supported by our study.

Kalifa et al. (2006) have investigated the relationship between focal drivers and sites of

fractionation using optical mapping in the experimental setting. In 8 isolated sheep

hearts they were able to show maximal DF regions in which fractionation was lowest

and an increased band of fractionation observed at border‐zones with areas of lower

DF in the posterior left atria. They therefore conclude that alongside areas of fast,

202

regular atrial activation, there are zones where propagation pattern variability and

fractionated activity occur most frequently. In the current study, high‐density mapping

in human atrial fibrillation identified clusters of high DF points, predominantly within

the left atrium, with fractionation observed at or adjacent to these sites. These

findings are consistent with data from the findings of Kalifa et al. (2006). These

investigators suggested that the success of the CFAE‐targeted ablation is due to

anatomical barriers being created to isolate the influence of high frequency “driver”

sites on the rest of the atria. The current clinical study suggests that such ablation in

the beating heart could additionally involve the closely positioned sites of high

activation frequency.

8.4.5 Clinical Implications

Activation mapping during atrial fibrillation has been notoriously difficult to perform in

the clinical setting. In addition, for longer duration atrial fibrillation, CFAE are observed

in many regions of the atria making identification of critical CFAE sites difficult. In this

setting, mapping activation frequency with DF may be of particular value in guiding

ablation in patients with more complex substrate such as in longstanding persistent

atrial fibrillation. A combined approach using both CFAE and DF mapping may allow a

more expeditious and targeted approach to ablation of the critical atrial substrate,

which further studies should address.

203

8.4.6 Limitations

Although this clinical study has provided high‐density contact mapping to evaluate the

relationship of CFAE and DF, its relative density does not approach that achieved by

optical mapping in the experimental setting. Higher density mapping may have allowed

identification of mechanisms leading to these phenomena. Nevertheless, our findings

on the relationship between CFAE and DF are similar to that achieved in such studies.

This study was unable to determine the effects of ablation on these sites due to the

off‐line nature of the analysis. In addition, the coronary sinus was not studied. Finally,

the questions of optimal sample length and the temporal stability of DF and CFE‐mean

over periods of minutes or more was not addressed by this study, but have been

studied elsewhere (Lazar et al. 2004; Sanders et al. 2005a; Lin et al. 2008; Stiles et al.

2008).

8.5 Conclusion

Greater fractionation and higher activation frequency are seen in persistent atrial

fibrillation and in the left atria. Significant regional variation in activation frequency is

observed in paroxysmal atrial fibrillation and this distinction diminishes in persistent

atrial fibrillation. CFAE and DF demonstrate no relationship on a point‐by‐point basis.

However, there is a relationship between CFAE and DF at an individual level. High‐

density exploration of their spatial relationship confirms the occurrence of CFAE

adjacent to areas of high frequency activation.

204

Table 8.1


Paroxysmal (n=10)

Persistent (n=10)

p value

Age (years) 61.5 55.0 0.1

Male, n (%) 6 (60) 10 (100) 0.09

Longest AF episode 3±2 days 15±7 months <0.0001

Time since first AF episode 51±37 months 59±53 months 0.7

Structural heart disease, n (%) 0 2 (20) 0.5

Amiodarone use, n (%) 1 (10) 5 (50) 0.1





LV end diastolic diameter (mm) 49±5 53±5 0.09

LV end systolic diameter (mm) 32±7 36±8 0.2

LV ejection fraction (%) 60±4 56±10 0.3

AF = atrial fibrillation, LV = left ventricular

205

Figure 8‐1 Integrated bi‐atrial high‐density CFE‐mean colour maps with example bipolar atrial

electrogram


206


The electrogram demonstrates yellow tick‐marks at the instant of maximum‐negative

slope on downstrokes between leading local‐maximum and trailing local‐minimum

points. Tick‐marks indicate annotations of deflection events in the electrogram

according to the Refractory, Peak‐to‐Peak Sensitivity and Duration criteria defined in

the text. CFE‐mean is the average time duration between annotations and the value

for each point is displayed along a colour spectrum on the 3‐dimensional geometry.

IVC – inferior vena cava , LAA – left atrial appendage, LIPV – left inferior pulmonary

vein, LSPV – left superior pulmonary vein, MA – mitral annulus, RAA – right atrial

appendage, RIPV – right inferior pulmonary vein, RSPV – right superior pulmonary vein,

SVC – superior vena cava, TA – tricuspid annulus.

207

Figure 8‐2

(a) Median CFE‐mean by atrial region

(b) Median DF by atrial region


208


(a) CFE‐mean was significantly lower in persistent compared to paroxysmal atrial

fibrillation for all regions (p<0.01). For post‐hoc pair‐wise comparisons regardless of

type of atrial fibrillation; *p<0.04 compared to all other regions, †p<0.05 compared to

LAA, ‡p<0.05 compared to any right atrial region, and §p<0.05 compared to LA Lat.

(b) All regions were significantly higher in persistent compared to paroxysmal atrial

fibrillation (p<0.01), except for RA Lat and RA Ant. Post‐hoc pair‐wise comparisons for

Paroxysmal atrial fibrillation; PVs, LAA and LA Roof were significantly higher than LA

Lat and LA Inf (p<0.05), and RA Ant was significantly higher than the 6 lowest areas.

Post‐hoc pair‐wise comparisons for Persistent atrial fibrillation; no significant

differences were seen between LA areas except LA Inf and the 3 highest areas

(p<0.05), and RA Lat, RA Post and RA Sept were significantly lower than many of the LA

areas and RA Ant (p<0.05).

PVs – pulmonary veins, LAA – left atrial appendage, LA – left atrial, RA – right atrial, Lat

– lateral wall, Post – posterior wall, Sept – septum, Ant – anterior wall, Inf – inferior

wall.

Figure 8‐3

Scatter plots of DF and CFE‐mean with “line of best fit”

Left pan

el: P

oint‐by‐po

int correlatio

n for every sample locatio

n in every patient.

Right p

anel: P

atient‐by‐patie

nt correlatio

n be

tween med

ians of C

FE‐m

ean and DF

210

Figure 8‐4

Antero‐superior view of the left atrium displaying DF data on a colour spectrum

High DF areas (blue‐purple) are seen in the pulmonary veins, basal left atrial

appendage and the left atrial roof. A zoomed view of the area of high DF at the roof is

shown with the corresponding points (inset). One second electrograms, frequency

spectra (3‐15Hz), and corresponding DF and CFE‐mean values for each of the local 7

points contributing to the colour map are shown in linked boxes. The central point of

highest DF is fast and regular, with surrounding points showing either increased

fractionation, lower DF or both.

211

Figure 8‐5

Integrated CT images with activation maps from the 5‐spline catheter at high DF cluster sites and accompanying bipolar electrograms


212


(a) Activation map at the left inferior pulmonary vein of a patient with paroxysmal

atrial fibrillation. Dominant frequency at this site was 7.5Hz (median DF 5.4Hz for this

left atrium). In this example, activation seems to radiate out from a distinctly early site

(A2). Note the increase in fractionation in Spline E with delayed conduction.

(b) Activation map at the left atrial roof of a patient with persistent atrial fibrillation.

Dominant frequency at this site was 8.4Hz (median DF 7.2Hz for this left atrium).

Earliest activation is seen in distal Spline B with Splines C and D activated soon after.

Delayed conduction is seen in the direction of Splines A and E with isochronal crowding

and fractionation.

(c) Activation map at the posterior left atrial wall of a patient with paroxysmal atrial

fibrillation. Dominant frequency at this site was 6.4Hz (median DF 5.8Hz for this left

atrium). Activation can be seen to move in an anti‐clockwise direction around the

splines with block between Splines A and E, evidenced by double potentials in Spline E.

213

Chapter 9. Summary‐‐‐‐‐

This thesis has examined aspects of the underlying substrate present in atrial

fibrillation and atrial flutter. The insights gained into the characteristics of the atrial

myocardium that initiate and sustain atrial arrhythmia have advanced our

understanding of the mechanisms responsible for this disease. This knowledge may

contribute to strategies aimed at prevention and treatment of atrial arrhythmia.

Studies of patients with recent atrial arrhythmia have demonstrated electrical and

structural remodelling within the atria. However, it is not known if this remodelling

precedes, occurs in association with, or results from the atrial arrhythmia. These

studies have suggested an important role for a decrease in atrial refractoriness and are

central to the observation that "atrial fibrillation begets atrial fibrillation". Clinical

studies have shown reversal of electrical remodelling over time after termination of

arrhythmia, however strategies of prompt termination of atrial fibrillation to avoid this

cycle of adverse remodelling have failed to show benefit. These observations have led

to the search for a “second factor” integral to the development and progression of

atrial fibrillation. Characterisation of the atrial substrate distant in time from episodes

of atrial fibrillation removes the effect that recent arrhythmia is known to have on

atrial electrophysiology. Chapter 2 has characterised the electrical substrate of the

right and left atria of patients with paroxysmal lone atrial fibrillation, remote from

recent arrhythmia, and demonstrated conduction abnormalities characterised by

prolongation of conduction times along linearly placed catheters, longer P wave

duration and site‐specific conduction delay; impaired sinus node function indicated by

prolonged SACT and CSNRT; and an increase in ERP, which is consistent with prior

214

studies evaluating clinical substrates for atrial fibrillation but is in contrast to the

remodelling attributed to atrial fibrillation itself. This suggests that patients with

paroxysmal lone atrial fibrillation have an abnormal atrial electrophysiological

substrate and that these abnormalities contribute to the “second factor” that

promotes progression of atrial fibrillation. Chapter 3 has characterised the

electroanatomical substrate present in lone atrial fibrillation. Several animal models

and clinical conditions known to be associated with atrial fibrillation have

demonstrated structural remodelling and conduction abnormalities without reduction

in refractoriness, but the changes observed may be influenced by the underlying

disease itself. Patients with “lone” atrial fibrillation do not have overt cardiovascular

disease to influence the underlying substrate. Therefore, electroanatomical

characterisation of the atria of patients with lone atrial fibrillation has aimed to

minimise the effects of associated disease and look at the substrate related to the

arrhythmia itself. In addition, this characterisation was performed remote from recent

episodes of arrhythmia to diminish the transient effects of rate‐related remodelling.

This study demonstrated bi‐atrial structural abnormalities characterised by atrial

dilatation and lower mean atrial voltage suggesting the loss of atrial myocardium;

slower conduction velocities; and an increase in the proportion of electrograms

characterised as double potentials or fractionated signals. These findings suggest that

patients with paroxysmal lone atrial fibrillation have an abnormal electroanatomical

atrial substrate and that these abnormalities also contribute to the “second factor”

that promotes atrial fibrillation.

The course of the typical atrial flutter circuit has been well characterised and ablation

of the cavotricuspid isthmus is a highly successful treatment. However, despite this

215

therapy a significant proportion subsequently develop atrial fibrillation. Indeed, atrial

flutter has a close inter‐relationship with atrial fibrillation and the two rhythms

frequently co‐exist. Chapter 4 postulates that causes for the propensity for future

atrial fibrillation in this population may be found by investigation of the underlying

electrical substrate of the right atrium in atrial flutter. Major findings from this study

include conduction abnormalities characterised by prolongation of lateral right atrial

conduction time, longer P wave duration and site‐specific conduction delay;

impairment of sinus node function; and an increase or no change in ERP. Thus, the

abnormal electrical substrate of atrial flutter is found to be present diffusely

throughout the right atrium and not confined to the barriers of the known typical

flutter circuit. Chapter 5 has characterised the electroanatomical substrate of atrial

flutter to further define abnormalities throughout the right atrium. Findings include

structural abnormalities reflected in atrial dilatation and lower mean right atrial

voltage suggesting the loss of atrial myocardium; conduction abnormalities in the form

of slower mean conduction velocity; and an increase in the proportion of complex

electrograms. Together with the electrical substrate determined in the prior chapter,

these changes reflect a diffusely abnormal right atrial substrate characteristic of the

changes seen in atrial fibrillation with prior studies and indeed earlier in this thesis.

Perhaps it is these abnormalities that predispose many patients with atrial flutter to

progress to atrial fibrillation.

Several investigators have observed remodelling of the sinus node that develops soon

after the initiation of arrhythmia and reverses after restoration of sinus rhythm. In

addition, it is known that several conditions predisposed to the development of atrial

arrhythmias demonstrate sinus node remodelling. Furthermore, atrial fibrillation not

216

uncommonly complicates sinus node disease. This frequent co‐existence of atrial

arrhythmia and sinus node disease may be due to a slow sinus rate increasing

vulnerability to atrial fibrillation, a common disease substrate and/or arrhythmia

directly impairing sinus node function. The anatomical sinus node is a small localised

structure located at the junction of the superior vena cava with the right atrium. In

contrast, the functional sinus node complex has been demonstrated to extend the

length of the crista terminalis and indeed beyond. This discrepancy is difficult to

reconcile and various hypotheses have been proposed. Newer mapping technology has

provided novel insights into arrhythmia mechanisms not previously afforded using

conventional techniques, demonstrating “preferential pathways” from earliest

activation to the site where break‐out occurs to the remaining atria in ectopic atrial

tachycardia. Whether similar pathways exist to account for the discrepancy between

the anatomical and functional sinus node complex is not known. Chapter 6 aimed to

characterise the electrical properties of the functional sinus node complex in the

normal heart with the anatomical and electrical resolution provided by a 3‐

dimensional non‐contact mapping system. To further explore the sinus node complex

activity, the same characterisation was performed in patients with chronic atrial flutter

(a group known to have impaired sinus node function) to establish how sinus node

activity is affected by atrial remodelling. This chapter found that the sinus node

complex in normal hearts displays a dynamic range of activation sites along the

postero‐lateral right atrium and preferential pathways of conduction do indeed exist

between the sinus node and the exit of sinus activity to the atria. There were multiple

origins of sinus activation and exit sites to the atria, providing evidence of multi‐

centricity of the sinus node complex. Lastly, remodelled atria demonstrate structural

217

change reflected by loss of voltage that was associated with more caudal activation,

slower conduction time along preferential pathways, only modest shifts of the

functional pacemaker complex and more frequent conduction block across the crista

terminalis with resultant circuitous conduction of the sinus impulse. These findings

extend our knowledge of the function of the highly complex human sinus node by

demonstrating the presence of preferential pathways of conduction within the

functional sinus node complex. The abnormalities in the sinus node complex shown for

the patients with atrial flutter underscore the important association between impaired

sinus node function and atrial arrhythmia.

As ablation techniques are used to treat patients with atrial fibrillation, it has become

apparent that many patients require not just trigger elimination but also substrate

modification. There is much work being done on identifying the appropriate atrial

substrate to be modified. Areas of CFAE and DF have been suggested as critical regions

maintaining atrial fibrillation. Techniques to modify the underlying atrial substrate by

ablation of these targets in order to treat atrial fibrillation have been proposed. CFAE‐

targeted ablation is one such technique and has been reported as highly successful,

although attempts to emulate the outcomes initially reported have shown more

modest improvements. Perhaps one reason for this non‐reproducibility is that the

evaluation of CFAE has been based on the operator’s subjective analysis. Mapping to

identify sites of high DF may identify the position of “drivers” of atrial fibrillation and

targeting these electrograms is a promising area of research. The development of

automated algorithms for both quantification of electrogram fractionation and real‐

time DF calculation aims to deliver a reproducible and objective assessment of

electrogram characteristics and thereby standardise the approach to substrate

218

modification. The ideal duration of electrograms required to reproducibly identify sites

of CFAE or highest DF is not known. While longer electrogram recordings extend the

data collection process unnecessarily, shorter durations may provide an inaccurate

representation of the true degree of fractionation or frequency. Chapter 7 aimed to

determine the minimum duration of electrogram recording that provides accurate

data without unduly lengthening the data collection process, by exploring the degree

of correlation between progressively shorter electrogram recordings with the longest

exportable data sample length. Recordings of shorter durations were compared to the

maximum duration recordable by the system using several criteria including mean

values, mean absolute differences, the proportion of points deviating >10% from the

index value and intra‐class correlation coefficients. Using these criteria, a recording

duration of less then 5 seconds was shown to disclose inaccurate data about the

fractionation and activation frequency of the electrograms reflecting underlying

substrate of atrial fibrillation. An electrogram duration of 5 seconds is therefore

recommended with these techniques of automated analysis.

The relationship between CFAE and sites of high DF has not been previously

demonstrated in humans. In particular, it is not known whether they are markers of

the same or provide complementary information on the pathogenesis of atrial

fibrillation. Understanding the relationship between these targets may be useful in

guiding radiofrequency ablation. In Chapter 8, the relationship between sites of high

frequency activation and fractionation during atrial fibrillation was examined by the

creation of high‐density bi‐atrial maps. This study found that the left atrium is more

highly fractionated than the right atrium and that patients with persistent atrial

fibrillation demonstrate a higher degree of fractionation than those with paroxysmal

219

atrial fibrillation. Similarly, activation frequency is higher in the left atrium than the

right atrium and higher in persistent atrial fibrillation than paroxysmal atrial

fibrillation. In paroxysmal atrial fibrillation regional analysis demonstrated significant

variation in frequency, but these regional differences were less apparent in persistent

atrial fibrillation. Correlation between fractionation and activation frequency exists but

is much stronger on a patient‐by‐patient rather than a point‐by‐point basis, prompting

the exploration of the spatial relationship between points of high DF and fractionation.

Detailed inspection of 3‐dimensional high‐density maps created during atrial

fibrillation reveals that electrogram fractionation is observed in close proximity to sites

of high frequency activation. This data is further supported by high‐density

simultaneous activation mapping at sites of high DF. The clinical implications of this

relationship are that mapping activation frequency with DF may be of particular value

in guiding ablation in patients with more complex substrate such as in longstanding

persistent atrial fibrillation. A combined approach using both CFAE and DF mapping

may allow a more expeditious and targeted approach to ablation of the critical atrial

substrate, which further studies should address.

220

Chapter 10. Future Directions‐‐‐

The characterisation of the underlying atrial substrate of atrial fibrillation and atrial

flutter performed for the purpose of this thesis has provided important insights into

the mechanism of the disease process associated with atrial arrhythmia. However,

many questions remain unanswered, some of which are discussed below.

This thesis provides documentation of the substrate of atrial fibrillation and of atrial

flutter under specific conditions. By design, the electrophysiological and

electroanatomical studies were performed remote from recent arrhythmia to minimise

the known transient effects of rate‐related remodelling. As such, the conclusions

drawn relate to the substrate present without the effects of recent arrhythmia which

of course may not reflect the clinical situation of many patients with atrial arrhythmia.

The complex interaction between the electrical and structural remodelling

demonstrated here and the rate‐related remodelling observed during and shortly after

arrhythmia requires further study.

Additionally, the similarly complex interaction between substrate for atrial fibrillation

and associated cardiovascular disease is an area for future investigation. The bi‐atrial

substrate identified for lone atrial fibrillation here necessarily excluded underlying

heart disease and most patients with atrial fibrillation do indeed have related disorders

which would variably impact on the atrial substrate.

The cause of atrial arrhythmia is widely recognised to be due to the presence of both

triggers and substrate to initiate and maintain the abnormal rhythm. This thesis has

concentrated on characterising the substrate underlying atrial arrhythmia. Although

221

extrapolation of data regarding sites of high DF as potential triggers sites could be

made, this was not the primary aim of the study. Therefore, detailed examination of

the electrophysiological characteristics of triggers of atrial arrhythmia is a potential

area for future research that would complement the work on atrial substrate.

Clinical studies such as those performed in this thesis make some assumptions about

disease process and observed findings – e.g. reduction in observed voltage is likely to

reflect loss of atrial myocardium. Insights into the precise mechanisms contributing to

the changes observed may come from animal studies where the methods by which

abnormalities manifest are more readily established.

A substantial body of work exists on the effect of the autonomic nervous system on

atrial fibrillation. This thesis did not set out to characterise the effect of the ganglionic

plexi on the atrial substrate and future work on the contribution of vagal influence on

the triggers and maintenance of atrial fibrillation is expected to increase our

knowledge in this important area.

Sinus node disease is a significant cause of morbidity and a leading indication for

permanent pacemaker implantation. The findings in this thesis of preferential

pathways being integral to the function of the sinus node complex, and the

observations of sinus node remodelling associated with atrial flutter being mediated

via these pathways, has implications for the understanding and perhaps eventual

management of this common condition.

Translating the findings regarding atrial substrate to a useful clinical intervention

would be a worthy aim. The studies of CFAE and DF have a potential role in the

effective guidance of ablation strategies aiming to modify the substrate of atrial

222

arrhythmia. Although the strategy of CFAE‐targeted ablation is well accepted, that of

targeting sites of high DF is not yet established. DF‐guided ablation is a potential direct

application of the data presented in Chapter 8 and future studies comparing standard

ablation techniques with or without additional targeting of sites of high DF may show a

difference in ablation efficacy. Furthermore, a combined approach (identifying both

CFAE and sites of high DF) of electrogram‐targeted ablation may utilise the

complementary information obtained from each technique to increase effective

ablation while decreasing unnecessary ablation of “bystander” myocardium not

actively involved in the substrate maintaining atrial fibrillation. As the indications for

atrial fibrillation ablation expand and the additional ablation required to maintain

outcomes in the more complex patient increases, a method that could reduce ablation

time and increase safety yet maintain efficacy would be welcome.

223

References

Aberg H. Atrial fibrillation. A review of 463 cases from Philadelphia General Hospital

from 1955 to 1965. Acta Med Scand 1968;184:425‐431.

Aimé‐Sempé C, Folliguet T, Rucker‐Martin C, Krajewska M, Krajewska S, Heimburger M,

Aubier M, Mercadier JJ, Reed JC and Hatem SN. Myocardial cell death in

fibrillating and dilated human right atria. J Am Coll Cardiol 1999;34:1577‐1586.

Alboni P, Malcarne C, Pedroni P, Masoni A and Narula OS. Electrophysiology of normal

sinus node with and without autonomic blockade. Circulation 1982;65:1236‐

1242.

Allessie M, Ausma J and Schotten U. Electrical, contractile and structural remodeling

during atrial fibrillation. Cardiovasc Res 2002;54:230‐246.

Allessie MA, Bonke FI and Schopman FJ. Circus movement in rabbit atrial muscle as a

mechanism of tachycardia. III. The "leading circle" concept: a new model of

circus movement in cardiac tissue without the involvement of an anatomical

obstacle. Circ Res 1977;41:9‐18.

Allessie MA, Boyden PA, Camm AJ, Kleber AG, Lab MJ, Legato MJ, Rosen MR, Schwartz

PJ, Spooner PM, Van Wagoner DR and Waldo AL. Pathophysiology and

prevention of atrial fibrillation. Circulation 2001;103:769‐777.

Anderson KR, Ho SY and Anderson RH. Location and vascular supply of sinus node in

human heart. Br Heart J 1979;41:28‐32.

Anderson RH, Ho SY, Smith A and Becker AE. The internodal atrial myocardium. Anat

Rec 1981a;201:75‐82.

224

Anderson RH, Becker AE, Tranum‐Jensen J and Janse MJ. Anatomico‐

electrophysiological correlations in the conduction system‐‐a review. Br Heart J

1981b;45:67‐82.

Arentz T, Haegeli L, Sanders P, Weber R, Neumann FJ, Kalusche D and Haïssaguerre M.

High‐density mapping of spontaneous pulmonary vein activity initiating atrial

fibrillation in humans. J Cardiovasc Electrophysiol 2007;18:31‐38.

Arora R, Verheule S, Scott L, Navarrete A, Katari V, Wilson E, Vaz D and Olgin JE.

Arrhythmogenic substrate of the pulmonary veins assessed by high‐resolution

optical mapping. Circulation 2003;107:1816‐1821.

Atienza F, Almendral J, Moreno J, Vaidyanathan R, Talkachou A, Kalifa J, Arenal A,

Villacastin JP, Torrecilla EG, Sanchez A, Ploutz‐Snyder R, Jalife J and Berenfeld

O. Activation of inward rectifier potassium channels accelerates atrial

fibrillation in humans: evidence for a reentrant mechanism. Circulation

2006;114:2434‐2442.

Ausma J, van der Velden HM, Lenders MH, van Ankeren EP, Jongsma HJ, Ramaekers FC,

Borgers M and Allessie MA. Reverse structural and gap‐junctional remodeling

after prolonged atrial fibrillation in the goat. Circulation 2003;107:2051‐2058.

Australian Institute of Health and Welfare, National Hospital Morbidity Database,

Separation statistics by principal diagnosis in ICD‐10‐AM, Australia, 1998‐99 to

2004‐2005. Retrieved 21 April 2008 from www.aihw.gov.au/cognos/cgi‐

bin/ppdscgi?DC=Q&E=/ahs/principaldiagnosis9899‐0405.

Badhwar N, Kalman JM, Sparks PB, Kistler PM, Attari M, Berger M, Lee RJ, Sra J and

Scheinman MM. Atrial tachycardia arising from the coronary sinus musculature:

225

electrophysiological characteristics and long‐term outcomes of radiofrequency

ablation. J Am Coll Cardiol 2005;46:1921‐1930.

Bailin SJ, Adler S and Giudici M. Prevention of chronic atrial fibrillation by pacing in the

region of Bachmann's bundle: results of a multicenter randomized trial. J

Cardiovasc Electrophysiol 2001;12:912‐917.

Becker AE. How structurally normal are human atria in patients with atrial fibrillation?

Heart Rhythm 2004;1:627‐631.

Benjamin EJ, Levy D, Vaziri SM, D'Agostino RB, Belanger AJ and Wolf PA. Independent

risk factors for atrial fibrillation in a population‐based cohort. The Framingham

Heart Study. JAMA 1994;271:840‐844.

Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB and Levy D. Impact of

atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation

1998;98:946‐952.

Berenfeld O, Mandapati R, Dixit S, Skanes AC, Chen J, Mansour M and Jalife J. Spatially

distributed dominant excitation frequencies reveal hidden organization in atrial

fibrillation in the Langendorff‐perfused sheep heart. J Cardiovasc Electrophysiol

2000;11:869‐879.

Betts TR, Roberts PR and Morgan JM. High‐density mapping of left atrial endocardial

activation during sinus rhythm and coronary sinus pacing in patients with

paroxysmal atrial fibrillation. J Cardiovasc Electrophysiol 2004;15:1111‐1117.

Blom NA, Gittenberger‐de Groot AC, DeRuiter MC, Poelmann RE, Mentink MM and

Ottenkamp J. Development of the cardiac conduction tissue in human embryos

using HNK‐1 antigen expression: possible relevance for understanding of

abnormal atrial automaticity. Circulation 1999;99:800‐806.

226

Blomstrom‐Lundqvist C, Scheinman MM, Aliot EM, Alpert JS, Calkins H, Camm AJ,

Campbell WB, Haines DE, Kuck KH, Lerman BB, Miller DD, Shaeffer CW,

Stevenson WG, Tomaselli GF, Antman EM, Smith SC, Jr., Alpert JS, Faxon DP,

Fuster V, Gibbons RJ, Gregoratos G, Hiratzka LF, Hunt SA, Jacobs AK, Russell RO,

Jr., Priori SG, Blanc JJ, Budaj A, Burgos EF, Cowie M, Deckers JW, Garcia MA,

Klein WW, Lekakis J, Lindahl B, Mazzotta G, Morais JC, Oto A, Smiseth O and

Trappe HJ. ACC/AHA/ESC guidelines for the management of patients with

supraventricular arrhythmias ‐ executive summary. A report of the American

College of Cardiology/American Heart Association task force on practice

guidelines and the European Society of Cardiology committee for practice

guidelines. J Am Coll Cardiol 2003;42:1493‐1531.

Boineau JP, Schuessler RB, Mooney CR, Wylds AC, Miller CB, Hudson RD, Borremans JM

and Brockus CW. Multicentric origin of the atrial depolarization wave: the

pacemaker complex. Relation to dynamics of atrial conduction, P‐wave changes

and heart rate control. Circulation 1978;58:1036‐1048.

Boineau JP, Schuessler RB, Hackel DB, Miller CB, Brockus CW and Wylds AC.

Widespread distribution and rate differentiation of the atrial pacemaker

complex. Am J Physiol 1980;239:H406‐415.

Boineau JP, Miller CB, Schuessler RB, Roeske WR, Autry LJ, Wylds AC and Hill DA.

Activation sequence and potential distribution maps demonstrating

multicentric atrial impulse origin in dogs. Circ Res 1984;54:332‐347.

Boineau JP, Canavan TE, Schuessler RB, Cain ME, Corr PB and Cox JL. Demonstration of

a widely distributed atrial pacemaker complex in the human heart. Circulation

1988;77:1221‐1237.

227

Boldt A, Scholl A, Garbade J, Resetar ME, Mohr FW, Gummert JF and Dhein S. ACE‐

inhibitor treatment attenuates atrial structural remodeling in patients with lone

chronic atrial fibrillation. Basic Res Cardiol 2006;101:261‐267.

Bouman LN, Gerlings ED, Biersteker PA and Bonke FI. Pacemaker shift in the sino‐atrial

node during vagal stimulation. Pflugers Arch 1968;302:255‐267.

Boyden PA, Tilley LP, Pham TD, Liu SK, Fenoglic JJ, Jr. and Wit AL. Effects of left atrial

enlargement on atrial transmembrane potentials and structure in dogs with

mitral valve fibrosis. Am J Cardiol 1982;49:1896‐1908.

Boyden PA, Tilley LP, Albala A, Liu SK, Fenoglio JJ, Jr. and Wit AL. Mechanisms for atrial

arrhythmias associated with cardiomyopathy: a study of feline hearts with

primary myocardial disease. Circulation 1984;69:1036‐1047.

Brand FN, Abbott RD, Kannel WB and Wolf PA. Characteristics and prognosis of lone

atrial fibrillation. 30‐year follow‐up in the Framingham Study. JAMA

1985;254:3449‐3453.

Braunwald E. Shattuck lecture‐‐cardiovascular medicine at the turn of the millennium:

triumphs, concerns, and opportunities. N Engl J Med 1997;337:1360‐1369.

Breithardt G, Seipel L and Loogen F. Sinus node recovery time and calculated sinoatrial

conduction time in normal subjects and patients with sinus node dysfunction.

Circulation 1977;56:43‐50.

Bromberg BI, Hand DE, Schuessler RB and Boineau JP. Primary negativity does not

predict dominant pacemaker location: implications for sinoatrial conduction.

Am J Physiol 1995;269:H877‐887.

Brown HF, DiFrancesco D and Noble SJ. How does adrenaline accelerate the heart?

Nature 1979;280:235‐236.

228

Brugada R, Tapscott T, Czernuszewicz GZ, Marian AJ, Iglesias A, Mont L, Brugada J,

Girona J, Domingo A, Bachinski LL and Roberts R. Identification of a genetic

locus for familial atrial fibrillation. N Engl J Med 1997;336:905‐911.

Cabrera JA, Ho SY, Climent V and Sanchez‐Quintana D. The architecture of the left

lateral atrial wall: a particular anatomic region with implications for ablation of

atrial fibrillation. Eur Heart J 2008;29:356‐362.

Cadilhac DA, Carter RC, Thrift AG and Dewey HM. Why invest in a national public

health program for stroke? An example using Australian data to estimate the

potential benefits and cost implications. Health Policy 2007;83:287‐294.

Calkins H, Brugada J, Packer DL, Cappato R, Chen SA, Crijns HJ, Damiano RJ, Jr., Davies

DW, Haines DE, Haïssaguerre M, Iesaka Y, Jackman W, Jaïs P, Kottkamp H, Kuck

KH, Lindsay BD, Marchlinski FE, McCarthy PM, Mont JL, Morady F, Nademanee

K, Natale A, Pappone C, Prystowsky E, Raviele A, Ruskin JN and Shemin RJ.

HRS/EHRA/ECAS expert Consensus Statement on catheter and surgical ablation

of atrial fibrillation: recommendations for personnel, policy, procedures and

follow‐up. A report of the Heart Rhythm Society (HRS) Task Force on catheter

and surgical ablation of atrial fibrillation. Heart Rhythm 2007;4:816‐861.

Callans DJ, Ren JF, Schwartzman D, Gottlieb CD, Chaudhry FA and Marchlinski FE.

Narrowing of the superior vena cava‐right atrium junction during

radiofrequency catheter ablation for inappropriate sinus tachycardia: analysis

with intracardiac echocardiography. J Am Coll Cardiol 1999;33:1667‐1670.

Cha YM, Redfield MM, Shen WK and Gersh BJ. Atrial fibrillation and ventricular

dysfunction: a vicious electromechanical cycle. Circulation 2004;109:2839‐

2843.

229

Chen CC, Tai CT, Chiang CE, Yu WC, Lee SH, Chen YJ, Hsieh MH, Tsai CF, Lee KW, Ding

YA, Chang MS and Chen SA. Atrial tachycardias originating from the atrial

septum: electrophysiologic characteristics and radiofrequency ablation. J


Chen SA, Chiang CE, Yang CJ, Cheng CC, Wu TJ, Wang SP, Chiang BN and Chang MS.

Sustained atrial tachycardia in adult patients. Electrophysiological

characteristics, pharmacological response, possible mechanisms, and effects of

radiofrequency ablation. Circulation 1994;90:1262‐1278.

Chen SA, Hsieh MH, Tai CT, Tsai CF, Prakash VS, Yu WC, Hsu TL, Ding YA and Chang MS.

Initiation of atrial fibrillation by ectopic beats originating from the pulmonary

veins: electrophysiological characteristics, pharmacological responses, and

effects of radiofrequency ablation. Circulation 1999;100:1879‐1886.

Chen YJ, Chen SA, Chen YC, Yeh HI, Chan P, Chang MS and Lin CI. Effects of rapid atrial

pacing on the arrhythmogenic activity of single cardiomyocytes from

pulmonary veins: implication in initiation of atrial fibrillation. Circulation

2001;104:2849‐2854.

Chinitz JS, Gerstenfeld EP, Marchlinski FE and Callans DJ. Atrial fibrillation is common

after ablation of isolated atrial flutter during long‐term follow‐up. Heart

Rhythm 2007;4:1029‐1033.

Chugh A, Latchamsetty R, Oral H, Elmouchi D, Tschopp D, Reich S, Igic P, Lemerand T,

Good E, Bogun F, Pelosi F, Jr. and Morady F. Characteristics of cavotricuspid

isthmus‐dependent atrial flutter after left atrial ablation of atrial fibrillation.

Circulation 2006;113:609‐615.

230

Chung MK, Martin DO, Sprecher D, Wazni O, Kanderian A, Carnes CA, Bauer JA, Tchou

PJ, Niebauer MJ, Natale A and Van Wagoner DR. C‐reactive protein elevation in

patients with atrial arrhythmias: inflammatory mechanisms and persistence of

atrial fibrillation. Circulation 2001;104:2886‐2891.

Ciaroni S, Cuenoud L and Bloch A. Clinical study to investigate the predictive

parameters for the onset of atrial fibrillation in patients with essential

hypertension. Am Heart J 2000;139:814‐819.

Cosio FG, Palacios J, Vidal JM, Cocina EG, Gomez‐Sanchez MA and Tamargo L.

Electrophysiologic studies in atrial fibrillation. Slow conduction of premature

impulses: a possible manifestation of the background for reentry. Am J Cardiol

1983;51:122‐130.

Cosio FG, Arribas F, Palacios J, Tascon J and Lopez‐Gil M. Fragmented electrograms and

continuous electrical activity in atrial flutter. Am J Cardiol 1986;57:1309‐1314.

Cosio FG, Arribas F, Barbero JM, Kallmeyer C and Goicolea A. Validation of double‐

spike electrograms as markers of conduction delay or block in atrial flutter. Am

J Cardiol 1988;61:775‐780.

Cosio FG, Lopez‐Gil M, Goicolea A, Arribas F and Barroso JL. Radiofrequency ablation of

the inferior vena cava‐tricuspid valve isthmus in common atrial flutter. Am J

Cardiol 1993;71:705‐709.

Coumel P. Paroxysmal atrial fibrillation: a disorder of autonomic tone? Eur Heart J

1994;15 Suppl A:9‐16.

Cox JL, Canavan TE, Schuessler RB, Cain ME, Lindsay BD, Stone C, Smith PK, Corr PB and

Boineau JP. The surgical treatment of atrial fibrillation. II. Intraoperative

231

electrophysiologic mapping and description of the electrophysiologic basis of

atrial flutter and atrial fibrillation. J Thorac Cardiovasc Surg 1991;101:406‐426.

Cox JL, Boineau JP, Schuessler RB, Kater KM and Lappas DG. Five‐year experience with

the maze procedure for atrial fibrillation. Ann Thorac Surg 1993;56:814‐823.

Cranefield PF. Action potentials, afterpotentials, and arrhythmias. Circulation Res

1977;41:415‐423.

Cushny AR and Edmunds CW. Paroxysmal irregularity of the heart and auricular

fibrillation. Am J Med Sci 1907;133.

Da Costa A, Romeyer C, Mourot S, Messier M, Cerisier A, Faure E and Isaaz K. Factors

associated with early atrial fibrillation after ablation of common atrial flutter. A

single centre prospective study. Eur Heart J 2002;23:498‐506.

Daoud EG, Weiss R, Augostini RS, Kalbfleisch SJ, Schroeder J, Polsinelli G and Hummel

JD. Remodeling of sinus node function after catheter ablation of right atrial

flutter. J Cardiovasc Electrophysiol 2002;13:20‐24.

Darbar D, Herron KJ, Ballew JD, Jahangir A, Gersh BJ, Shen WK, Hammill SC, Packer DL

and Olson TM. Familial atrial fibrillation is a genetically heterogeneous disorder.

J Am Coll Cardiol 2003;41:2185‐2192.

Davies MJ and Pomerance A. Pathology of atrial fibrillation in man. Br Heart J

1972;34:520‐525.

De Ponti R, Ho SY, Salerno‐Uriarte JA, Tritto M and Spadacini G. Electroanatomic

analysis of sinus impulse propagation in normal human atria. J Cardiovasc

Electrophysiol 2002;13:1‐10.

de Voogt W, van Hemel N, de Vusser P, Mairesse GH, van Mechelen R, Koistinen J, van

den Bos A, Roose I, Voitk J, Yli‐Mayry S, Stockman D, El Allaf D, Tse HF and Lau

232

CP. No evidence of automatic atrial overdrive pacing efficacy on reduction of

paroxysmal atrial fibrillation. Europace 2007;9:798‐804.

Earley MJ, Abrams DJ, Sporton SC and Schilling RJ. Validation of the noncontact

mapping system in the left atrium during permanent atrial fibrillation and sinus

rhythm. J Am Coll Cardiol 2006;48:485‐491.

Eikelboom JW and Hankey GJ. The beginning of the end of warfarin? Med J Aust

2004;180:549‐551.

Einthoven W. Le telecardiogramme. Arch Int Physiol 1906;4:132‐164.

Ellinor PT, Yoerger DM, Ruskin JN and Macrae CA. Familial aggregation in lone atrial

fibrillation. Hum Genet 2005;118:179‐184.

Ellis K, Wazni O, Marrouche N, Martin D, Gillinov M, McCarthy P, Saad EB, Bhargava M,

Schweikert R, Saliba W, Bash D, Rossillo A, Erciyes D, Tchou P and Natale A.

Incidence of atrial fibrillation post‐cavotricuspid isthmus ablation in patients

with typical atrial flutter: left‐atrial size as an independent predictor of atrial

fibrillation recurrence. J Cardiovasc Electrophysiol 2007;18:799‐802.

Elvan A, Wylie K and Zipes DP. Pacing‐induced chronic atrial fibrillation impairs sinus

node function in dogs. Electrophysiological remodeling. Circulation

1996;94:2953‐2960.

Ernestene AC and Levine SA. A comparison of records taken with the Einthoven string

galvanometer and the amplifier type electrocardiograph. Am Heart J

1928;4:725‐731.

Estes NA, 3rd, Halperin JL, Calkins H, Ezekowitz MD, Gitman P, Go AS, McNamara RL,

Messer JV, Ritchie JL, Romeo SJ, Waldo AL, Wyse DG, Bonow RO, DeLong E,

Goff DC, Jr., Grady K, Green LA, Hiniker A, Linderbaum JA, Masoudi FA, Pina IL,

233

Pressler S, Radford MJ and Rumsfeld JS. ACC/AHA/Physician Consortium 2008

Clinical performance measures for adults with nonvalvular atrial fibrillation or

atrial flutter: A report of the American College of Cardiology/American Heart

Association task force on performance measures and the physician consortium

for performance improvement. J Am Coll Cardiol 2008;51:865‐884.

Everett TH, Wilson EE, Verheule S, Guerra JM, Foreman S and Olgin JE. Structural atrial

remodeling alters the substrate and spatiotemporal organization of atrial

fibrillation: a comparison in canine models of structural and electrical atrial

remodeling. Am J Physiol Heart Circ Physiol 2006;291:H2911‐H2923.

Fan K, Lee KL, Chow WH, Chau E and Lau CP. Internal cardioversion of chronic atrial

fibrillation during percutaneous mitral commissurotomy: insight into reversal of

chronic stretch‐induced atrial remodeling. Circulation 2002;105:2746‐2752.

Fareh S, Villemaire C and Nattel S. Importance of refractoriness heterogeneity in the

enhanced vulnerability to atrial fibrillation induction caused by tachycardia‐

induced atrial electrical remodeling. Circulation 1998;98:2202‐2209.

Feld GK, Fleck RP, Chen PS, Boyce K, Bahnson TD, Stein JB, Calisi CM and Ibarra M.

Radiofrequency catheter ablation for the treatment of human type 1 atrial

flutter. Identification of a critical zone in the reentrant circuit by endocardial

mapping techniques. Circulation 1992;86:1233‐1240.

Ferrer MI. The sick sinus syndrome in atrial disease. JAMA 1968;206:645‐646.

Ferrier GR and Moe GK. Effect of calcium on acetylstrophanthidin‐induced transient

depolarizations in canine Purkinje tissue. Circulation Res 1973;33:508‐515.

Flegel KM. From Delirium Cordis to Atrial Fibrillation: Historical Development of a

Disease Concept. Ann Intern Med 1995;122:867‐873.

234

Frame LH, Page RL and Hoffman BF. Atrial reentry around an anatomic barrier with a

partially refractory excitable gap. A canine model of atrial flutter. Circulation

Res 1986;58:495‐511.

Franz MR, Karasik PL, Li C, Moubarak J and Chavez M. Electrical remodeling of the

human atrium: similar effects in patients with chronic atrial fibrillation and

atrial flutter. J Am Coll Cardiol 1997;30:1785‐1792.

Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA and Maseri A. Histological

substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation

1997;96:1180‐1184.

Furberg CD, Psaty BM, Manolio TA, Gardin JM, Smith VE and Rautaharju PM.

Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health

Study). Am J Cardiol 1994;74:236‐241.

Fuster V, Ryden LE, Asinger RW, Cannom DS, Crijns HJ, Frye RL, Halperin JL, Kay GN,

Klein WW, Levy S, McNamara RL, Prystowsky EN, Wann LS and Wyse DG.

ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation.

A report of the American College of Cardiology/American Heart Association

Task Force on Practice Guidelines and the European Society of Cardiology

Committee for Practice Guidelines and Policy Conferences. Eur Heart J

2001;22:1852‐1923.

Fuster V, Ryden LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, Halperin JL, Le

Heuzey JY, Kay GN, Lowe JE, Olsson SB, Prystowsky EN, Tamargo JL, Wann S,

Smith SC, Jr., Jacobs AK, Adams CD, Anderson JL, Antman EM, Halperin JL, Hunt

SA, Nishimura R, Ornato JP, Page RL, Riegel B, Priori SG, Blanc JJ, Budaj A, Camm

AJ, Dean V, Deckers JW, Despres C, Dickstein K, Lekakis J, McGregor K, Metra

235

M, Morais J, Osterspey A, Tamargo JL and Zamorano JL. ACC/AHA/ESC 2006

Guidelines for the Management of Patients with Atrial Fibrillation: a report of

the American College of Cardiology/American Heart Association Task Force on

Practice Guidelines and the European Society of Cardiology Committee for

Practice Guidelines. Circulation 2006;114:e257‐354.

Fynn SP, Todd DM, Hobbs WJ, Armstrong KL, Fitzpatrick AP and Garratt CJ. Clinical

evaluation of a policy of early repeated internal cardioversion for recurrence of

atrial fibrillation. J Cardiovasc Electrophysiol 2002;13:135‐141.

Garratt CJ, Duytschaever M, Killian M, Dorland R, Mast F and Allessie MA. Repetitive

electrical remodeling by paroxysms of atrial fibrillation in the goat: no

cumulative effect on inducibility or stability of atrial fibrillation. J Cardiovasc


Garrey WE. The nature of fibrillary contraction of the heart: its relation to tissue mass

and form. Am J Physiol 1914;33:397‐414.

Gaspo R, Bosch RF, Talajic M and Nattel S. Functional mechanisms underlying

tachycardia‐induced sustained atrial fibrillation in a chronic dog model.

Circulation 1997;96:4027‐4035.

Geesbreght JM and Randall WC. Area localization of shifting cardiac pacemakers during

sympathetic stimulation. Am J Physiol 1971;220:1522‐1527.

Gepstein L, Hayam G and Ben‐Haim SA. A novel method for nonfluoroscopic catheter‐

based electroanatomical mapping of the heart. In vitro and in vivo accuracy

results. Circulation 1997;95:1611‐1622.

Gerstenfeld EP, Sahakian AV and Swiryn S. Evidence for transient linking of atrial

excitation during atrial fibrillation in humans. Circulation 1992;86:375‐382.

236

Gerstenfeld EP, Callans DJ, Sauer W, Jacobson J and Marchlinski FE. Reentrant and

nonreentrant focal left atrial tachycardias occur after pulmonary vein isolation.

Heart Rhythm 2005;2:1195‐1202.

Goette A, Honeycutt C and Langberg JJ. Electrical remodeling in atrial fibrillation. Time

course and mechanisms. Circulation 1996;94:2968‐2974.

Goldberg JM. Intra‐SA‐nodal pacemaker shifts induced by autonomic nerve stimulation

in the dog. Am J Physiol 1975;229:1116‐1123.

Goldberg JM, Lynn‐Johnson MH and Neely B. Use of P wave morphology for inferring

pacemaker localization along the sulcus terminalis in the dog. J Electrocardiol

1981;14:115‐124.

Goldstein RN, Ryu K, Khrestian C, van Wagoner DR and Waldo AL. Prednisone prevents

inducible atrial flutter in the canine sterile pericarditis model. J Cardiovasc


Gomes JA, Kang PS, Matheson M, Gough WB, Jr. and El‐Sherif N. Coexistence of sick

sinus rhythm and atrial flutter‐fibrillation. Circulation 1981;63:80‐86.

Goto M, Sakamoto Y and Imanaga I (1967). Aconitine‐induced fibrillation of the

different muscle tissues of the heart and the action of acetylcholine.

Electrophysiology and Ultrastructure of the Heart. T Sano, K Matsuda and B

Mizuhira. New York, Grune & Stratton: 199‐201.

Granada J, Uribe W, Chyou PH, Maassen K, Vierkant R, Smith PN, Hayes J, Eaker E and

Vidaillet H. Incidence and predictors of atrial flutter in the general population. J

Am Coll Cardiol 2000;36:2242‐2246.

Grigioni F, Avierinos JF, Ling LH, Scott CG, Bailey KR, Tajik AJ, Frye RL and Enriquez‐

Sarano M. Atrial fibrillation complicating the course of degenerative mitral

237

regurgitation: determinants and long‐term outcome. J Am Coll Cardiol

2002;40:84‐92.

Hachiya H, Ernst S, Ouyang F, Mavrakis H, Chun J, Bansch D, Antz M and Kuck KH.

Topographic distribution of focal left atrial tachycardias defined by

electrocardiographic and electrophysiological data. Circ J 2005;69:205‐210.

Hadian D, Zipes DP, Olgin JE and Miller JM. Short‐term rapid atrial pacing produces

electrical remodeling of sinus node function in humans. J Cardiovasc


Haïssaguerre M, Fischer B, Labbe T, Lemetayer P, Montserrat P, d'Ivernois C, Dartigues

JF and Warin JF. Frequency of recurrent atrial fibrillation after catheter ablation

of overt accessory pathways. Am J Cardiol 1992;69:493‐497.

Haïssaguerre M, Jaïs P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le MA,

Le MP and Clementy J. Spontaneous initiation of atrial fibrillation by ectopic

beats originating in the pulmonary veins. N Engl J Med 1998;339:659‐666.

Haïssaguerre M, Shah DC, Jaïs P, Hocini M, Yamane T, Deisenhofer I, Garrigue S and

Clementy J. Mapping‐guided ablation of pulmonary veins to cure atrial

fibrillation. Am J Cardiol 2000a;86:9K‐19K.

Haïssaguerre M, Shah DC, Jaïs P, Hocini M, Yamane T, Deisenhofer I, Chauvin M,

Garrigue S and Clementy J. Electrophysiological breakthroughs from the left

atrium to the pulmonary veins. Circulation 2000b;102:2463‐2465.

Haïssaguerre M, Jaïs P, Shah DC, Arentz T, Kalusche D, Takahashi A, Garrigue S, Hocini

M, Peng JT and Clementy J. Catheter ablation of chronic atrial fibrillation

targeting the reinitiating triggers. J Cardiovasc Electrophysiol 2000c;11:2‐10.

238

Haïssaguerre M, Sanders P, Hocini M, Hsu LF, Shah DC, Scavee C, Takahashi Y, Rotter

M, Pasquie JL, Garrigue S, Clementy J and Jaïs P. Changes in atrial fibrillation

cycle length and inducibility during catheter ablation and their relation to

outcome. Circulation 2004a;109:3007‐3013.

Haïssaguerre M, Sanders P, Hocini M, Jaïs P and Clementy J. Pulmonary veins in the

substrate for atrial fibrillation: the "venous wave" hypothesis. J Am Coll Cardiol

2004b;43:2290‐2292.

Haïssaguerre M, Sanders P, Hocini M, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu

LF, Bordachar P, Reuter S, Roudaut R, Clementy J and Jaïs P. Catheter ablation

of long‐lasting persistent atrial fibrillation: critical structures for termination. J

Cardiovasc Electrophysiol 2005a;16:1125‐1137.

Haïssaguerre M, Hocini M, Sanders P, Sacher F, Rotter M, Takahashi Y, Rostock T, Hsu


of long‐lasting persistent atrial fibrillation: clinical outcome and mechanisms of

subsequent arrhythmias. J Cardiovasc Electrophysiol 2005b;16:1138‐1147.

Haïssaguerre M, Sanders P, Hocini M, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu


of long‐lasting persistent atrial fibrillation: critical structures for termination. J

Cardiovasc Electrophysiol 2005c;16:1125‐1137.

Haïssaguerre M, Hocini M, Sanders P, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu

LF, Jonsson A, O'Neill MD, Bordachar P, Reuter S, Roudaut R, Clementy J and

Jaïs P. Localized sources maintaining atrial fibrillation organized by prior

ablation. Circulation 2006a;113:616‐625.

239

Haïssaguerre M, Hocini M, Sanders P, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu

LF, Jonsson A, O'Neill MD, Bordachar P, Reuter S, Roudaut R, Clementy J and

Jaïs P. Localized sources maintaining atrial fibrillation organized by prior

ablation. Circulation 2006b;113:616‐625.

Halligan SC, Gersh BJ, Brown RD, Jr., Rosales AG, Munger TM, Shen WK, Hammill SC

and Friedman PA. The natural history of lone atrial flutter. Ann Intern Med

2004;140:265‐268.

Hariman RJ, Hoffman BF and Naylor RE. Electrical activity from the sinus node region in

conscious dogs. Circ Res 1980;47:775‐791.

Hermida JS, Kubala M, Lescure FX, Delonca J, Clerc J, Otmani A, Jarry G and Rey JL.

Atrial septal pacing to prevent atrial fibrillation in patients with sinus node

dysfunction: results of a randomized controlled study. Am Heart J

2004;148:312‐317.

Higa S, Tai CT, Lin YJ, Liu TY, Lee PC, Huang JL, Hsieh MH, Yuniadi Y, Huang BH, Lee SH,

Ueng KC, Ding YA and Chen SA. Focal atrial tachycardia: new insight from

noncontact mapping and catheter ablation. Circulation 2004;109:84‐91.

Hobbs WJ, Fynn S, Todd DM, Wolfson P, Galloway M and Garratt CJ. Reversal of atrial

electrical remodeling after cardioversion of persistent atrial fibrillation in

humans. Circulation 2000;101:1145‐1151.

Hocini M, Ho SY, Kawara T, Linnenbank AC, Potse M, Shah D, Jaïs P, Janse MJ,

Haïssaguerre M and De Bakker JM. Electrical conduction in canine pulmonary

veins: electrophysiological and anatomic correlation. Circulation

2002;105:2442‐2448.

240

Hocini M, Sanders P, Deisenhofer I, Jaïs P, Hsu LF, Scavee C, Weerasoriya R, Raybaud F,

Macle L, Shah DC, Garrigue S, Le MP, Clementy J and Haïssaguerre M. Reverse

remodeling of sinus node function after catheter ablation of atrial fibrillation in

patients with prolonged sinus pauses. Circulation 2003;108:1172‐1175.

Hoff HE and Geddes LA. Cholinergic factor in auricular fibrillation. J Appl Physiol

1955;8:177‐192.

Hsieh MH, Tai CT, Chiang CE, Tsai CF, Yu WC, Chen YJ, Ding YA and Chen SA. Recurrent

atrial flutter and atrial fibrillation after catheter ablation of the cavotricuspid

isthmus: a very long‐term follow‐up of 333 patients. J Interv Card Electrophysiol

2002;7:225‐231.

Huang JL, Tai CT, Lin YJ, Ueng KC, Huang BH, Lee KT, Higa S, Yuniadi Y, Chang SL, Lo LW,

Wongcharoen W, Hu YF, Lee PC, Tuan TC, Ting CT and Chen SA. Right atrial

substrate properties associated with age in patients with typical atrial flutter.

Heart Rhythm 2008;5:1144‐1151.

Inoue H, Matsuo H, Takayanagi K and Murao S. Clinical and experimental studies of the

effects of extrastimulation and rapid pacing on atrial flutter: evidence of

macroreentry with an excitable gap. Am J Cardiol 1981;48:623‐631.

Jahangir A, Lee V, Friedman PA, Trusty JM, Hodge DO, Kopecky SL, Packer DL, Hammill

SC, Shen WK and Gersh BJ. Long‐term progression and outcomes with aging in

patients with lone atrial fibrillation: a 30‐year follow‐up study. Circulation

2007;115:3050‐3056.

Jaïs P, Haïssaguerre M, Shah DC, Chouairi S and Clementy J. Regional disparities of

endocardial atrial activation in paroxysmal atrial fibrillation. Pacing Clin


241

Jaïs P, Peng JT, Shah DC, Garrigue S, Hocini M, Yamane T, Haïssaguerre M, Barold SS,

Roudaut R and Clementy J. Left ventricular diastolic dysfunction in patients with

so‐called lone atrial fibrillation. J Cardiovasc Electrophysiol 2000;11:623‐625.

Jaïs P, Hocini M, Macle L, Choi KJ, Deisenhofer I, Weerasooriya R, Shah DC, Garrigue S,

Raybaud F, Scavee C, le MP, Clementy J and Haïssaguerre M. Distinctive

electrophysiological properties of pulmonary veins in patients with atrial

fibrillation. Circulation 2002;106:2479‐2485.

Jalife J, Berenfeld O, Skanes A and Mandapati R. Mechanisms of atrial fibrillation:

mother rotors or multiple daughter wavelets, or both? J Cardiovasc

Electrophysiol 1998;9:S2‐12.

Jalife J, Berenfeld O and Mansour M. Mother rotors and fibrillatory conduction: a

mechanism of atrial fibrillation. Cardiovasc Res 2002;54:204‐216.

James TN, Sherf L, Fine G and Morales AR. Comparative ultrastructure of the sinus

node in man and dog. Circulation 1966;34:139‐163.

James TN. The sinus node. Am J Cardiol 1977;40:965‐986.

John B, Stiles MK, Kuklik P, Chandy ST, Young GD, Mackenzie L, Szumowski L, Joseph G,

Worthley SG, Kalman JM and Sanders P. Electrical remodeling of the left and

right atria due to rheumatic mitral stenosis. Eur Heart J online

early;doi:10.1093/eurheartj/ehn1329.

Jolly WA and Ritchie TW. Auricular flutter and fibrillation. Heart 1911;3:177‐221.

Jorgensen HS, Nakayama H, Reith J, Raaschou HO and Olsen TS. Acute stroke with

atrial fibrillation. The Copenhagen Stroke Study. Stroke 1996;27:1765‐1769.

242

Kadish A, Hauck J, Pederson B, Beatty G and Gornick C. Mapping of atrial activation

with a noncontact, multielectrode catheter in dogs. Circulation 1999;99:1906‐

1913.

Kalifa J, Tanaka K, Zaitsev AV, Warren M, Vaidyanathan R, Auerbach D, Pandit S,

Vikstrom KL, Ploutz‐Snyder R, Talkachou A, Atienza F, Guiraudon G, Jalife J and

Berenfeld O. Mechanisms of wave fractionation at boundaries of high‐

frequency excitation in the posterior left atrium of the isolated sheep heart

during atrial fibrillation. Circulation 2006;113:626‐633.

Kalman JM, Lee RJ, Fisher WG, Chin MC, Ursell P, Stillson CA, Lesh MD and Scheinman

MM. Radiofrequency catheter modification of sinus pacemaker function guided

by intracardiac echocardiography. Circulation 1995;92:3070‐3081.

Kalman JM, Olgin JE, Saxon LA, Fisher WG, Lee RJ and Lesh MD. Activation and

entrainment mapping defines the tricuspid annulus as the anterior barrier in

typical atrial flutter. Circulation 1996;94:398‐406.

Kalman JM, Olgin JE, Saxon LA, Lee RJ, Scheinman MM and Lesh MD.

Electrocardiographic and electrophysiologic characterization of atypical atrial

flutter in man: use of activation and entrainment mapping and implications for

catheter ablation. J Cardiovasc Electrophysiol 1997;8:121‐144.

Kalman JM, Olgin JE, Karch MR, Hamdan M, Lee RJ and Lesh MD. "Cristal tachycardias":

origin of right atrial tachycardias from the crista terminalis identified by

intracardiac echocardiography. J Am Coll Cardiol 1998;31:451‐459.

Kannel WB, Abbott RD, Savage DD and McNamara PM. Epidemiologic features of

chronic atrial fibrillation: the Framingham study. N Engl J Med 1982a;306:1018‐

1022.

243

Kannel WB, Abbott RD, Savage DD and McNamara PM. Epidemiologic features of

chronic atrial fibrillation: the Framingham study. N Engl J Med 1982b;306:1018‐

1022.

Karagueuzian HS, Athill CA, Yashima M, Ikeda T, Wu TJ, Mandel WJ and Chen PS.

Transmembrane potential properties of atrial cells at different sites of a spiral

wave reentry: cellular evidence for an excitable but nonexcited core. Pacing

Clin Electrophysiol 1998;21:2360‐2365.

Kistler PM, Sanders P, Fynn SP, Stevenson IH, Hussin A, Vohra JK, Sparks PB and Kalman

JM. Electrophysiological and electrocardiographic characteristics of focal atrial

tachycardia originating from the pulmonary veins: acute and long‐term

outcomes of radiofrequency ablation. Circulation 2003a;108:1968‐1975.

Kistler PM, Sanders P, Hussin A, Morton JB, Vohra JK, Sparks PB and Kalman JM. Focal

atrial tachycardia arising from the mitral annulus: electrocardiographic and

electrophysiologic characterization. J Am Coll Cardiol 2003b;41:2212‐2219.

Kistler PM, Sanders P, Fynn SP, Stevenson IH, Spence SJ, Vohra JK, Sparks PB and

Kalman JM. Electrophysiologic and electroanatomic changes in the human

atrium associated with age. J Am Coll Cardiol 2004;44:109‐116.

Kistler PM, Fynn SP, Haqqani H, Stevenson IH, Vohra JK, Morton JB, Sparks PB and

Kalman JM. Focal atrial tachycardia from the ostium of the coronary sinus:

electrocardiographic and electrophysiological characterization and

radiofrequency ablation. J Am Coll Cardiol 2005;45:1488‐1493.

Kistler PM, Sanders P, Dodic M, Spence SJ, Samuel CS, Zhao C, Charles JA, Edwards GA

and Kalman JM. Atrial electrical and structural abnormalities in an ovine model

of chronic blood pressure elevation after prenatal corticosteroid exposure:

244

implications for development of atrial fibrillation. Eur Heart J 2006;27:3045‐

3056.

Klein GJ, Guiraudon GM, Sharma AD and Milstein S. Demonstration of macroreentry

and feasibility of operative therapy in the common type of atrial flutter. Am J

Cardiol 1986;57:587‐591.

Klos M, Calvo D, Yamazaki M, Zlochiver S, Mironov S, Cabrera J‐A, Sanchez‐Quintana D,

Jalife J, Berenfeld O and Kalifa J. The atrial septopulmonary bundle of the

posterior left atrium provides a substrate for AF initiation in a model of vagally

mediated pulmonary vein tachycardia of the structurally normal heart. Circ

Arrhythmia Electrophysiol 2008;1:175‐183.

Kojodjojo P, Peters NS, Davies DW and Kanagaratnam P. Characterization of the

electroanatomical substrate in human atrial fibrillation: the relationship

between changes in atrial volume, refractoriness, wavefront propagation

velocities, and AF burden. J Cardiovasc Electrophysiol 2007;18:269‐275.

Konings KT, Kirchhof CJ, Smeets JR, Wellens HJ, Penn OC and Allessie MA. High‐density

mapping of electrically induced atrial fibrillation in humans. Circulation

1994;89:1665‐1680.

Konings KT, Smeets JL, Penn OC, Wellens HJ and Allessie MA. Configuration of unipolar

atrial electrograms during electrically induced atrial fibrillation in humans.

Circulation 1997;95:1231‐1241.

Kopecky SL, Gersh BJ, McGoon MD, Whisnant JP, Holmes DR, Jr., Ilstrup DM and Frye

RL. The natural history of lone atrial fibrillation. A population‐based study over

three decades. N Engl J Med 1987;317:669‐674.

245

Kostin S, Klein G, Szalay Z, Hein S, Bauer EP and Schaper J. Structural correlate of atrial

fibrillation in human patients. Cardiovasc Res 2002;54:361‐379.

Kroll M, Kriebel T, Windhagen‐Mahnert B, Franzbach B, Jux C, Zutz M, Tebbenjohanns J

and Paul T. Origin of electrical activation within the right atrial and left

ventricular walls: differentiation by electrogram characteristics using the

noncontact mapping system. Pacing Clin Electrophysiol 2003;26:1970‐1978.

Kuklik P, Szumowski L, Zebrowski JJ and Walczak F. The reconstruction, from a set of

points, and analysis of the interior surface of the heart chamber. Physiol Meas

2004;25:617‐627.

Kulbertus HE, Leval‐Rutten F, Mary L and Casters P. Sinus node recovery time in the

elderly. Br Heart J 1975;37:420‐425.

Kumagai K, Akimitsu S, Kawahira K, Kawanami F, Yamanouchi Y, Hiroki T and Arakawa

K. Electrophysiological properties in chronic lone atrial fibrillation. Circulation

1991;84:1662‐1668.

Kumagai K, Ogawa M, Noguchi H, Yasuda T, Nakashima H and Saku K.

Electrophysiologic properties of pulmonary veins assessed using a

multielectrode basket catheter. J Am Coll Cardiol 2004;43:2281‐2289.

Kwong KF, Schuessler RB, Green KG, Laing JG, Beyer EC, Boineau JP and Saffitz JE.

Differential expression of gap junction proteins in the canine sinus node.

Circulation Res 1998;82:604‐612.

Lai LP, Lin JL and Huang SK. Molecular genetic studies in atrial fibrillation. Cardiology

2003;100:109‐113.

Lake FR, Cullen KJ, de Klerk NH, McCall MG and Rosman DL. Atrial fibrillation and

mortality in an elderly population. Aust N Z J Med 1989;19:321‐326.

246

Lau CP, Tse HF and Ayers GM. Defibrillation‐guided radiofrequency ablation of atrial

fibrillation secondary to an atrial focus. J Am Coll Cardiol 1999;33:1217‐1226.

Lazar S, Dixit S, Marchlinski FE, Callans DJ and Gerstenfeld EP. Presence of left‐to‐right

atrial frequency gradient in paroxysmal but not persistent atrial fibrillation in


Lazar S, Dixit S, Callans DJ, Lin D, Marchlinski FE and Gerstenfeld EP. Effect of

pulmonary vein isolation on the left‐to‐right atrial dominant frequency gradient

in human atrial fibrillation. Heart Rhythm 2006;3:889‐895.

Lee RJ, Kalman JM, Fitzpatrick AP, Epstein LM, Fisher WG, Olgin JE, Lesh MD and

Scheinman MM. Radiofrequency catheter modification of the sinus node for

"inappropriate" sinus tachycardia. Circulation 1995;92:2919‐2928.

Lee SH, Lin FY, Yu WC, Cheng JJ, Kuan P, Hung CR, Chang MS and Chen SA. Regional

differences in the recovery course of tachycardia‐induced changes of atrial

electrophysiological properties. Circulation 1999;99:1255‐1264.

Lemery R, Birnie D, Tang AS, Green M and Gollob M. Feasibility study of endocardial

mapping of ganglionated plexuses during catheter ablation of atrial fibrillation.

Heart Rhythm 2006;3:387‐396.

Lemery R, Birnie D, Tang AS, Green M, Gollob M, Hendry M and Lau E. Normal atrial

activation and voltage during sinus rhythm in the human heart: an endocardial

and epicardial mapping study in patients with a history of atrial fibrillation. J


Lemola K, Ting M, Gupta P, Anker JN, Chugh A, Good E, Reich S, Tschopp D, Igic P,

Elmouchi D, Jongnarangsin K, Bogun F, Pelosi F, Jr., Morady F and Oral H.

247

Effects of two different catheter ablation techniques on spectral characteristics

of atrial fibrillation. J Am Coll Cardiol 2006;48:340‐348.

Lewis T. Auricular fibrillation and its relationship to clinical irregularity of the heart.

1910;1:306‐372.

Lewis T, Oppenheimer BS and Oppenheimer A. The site of origin of the mammalian

heart beat: The pacemaker in the dog. Heart 1910;2:147‐169.

Lewis T and Schleiter HG. The relation of regular tachycardias of auricular origin to

auricular fibrillation. Heart 1912;3:173‐193.

Lewis T. Observations upon a curious, not uncommon form of extreme acceleration of

the auricle; atrial flutter. Heart 1913;4.

Lewis T, Drury AN and Bulger HA. Observations upon atrial flutter and fibrillation VI.

Heart 1921a;8:84‐134.

Lewis T, Drury AN and Iliescu CC. A demonstration of circus movement in clinical

fibrillation of the auricles. Heart 1921b;8:361‐369.

Li D, Fareh S, Leung TK and Nattel S. Promotion of atrial fibrillation by heart failure in

dogs: atrial remodeling of a different sort. Circulation 1999;100:87‐95.

Li D, Melnyk P, Feng J, Wang Z, Petrecca K, Shrier A and Nattel S. Effects of

experimental heart failure on atrial cellular and ionic electrophysiology.

Circulation 2000;101:2631‐2638.

Liebman J. Are there internodal tracts? Yes. Int J Cardiol 1985;7:175‐185.

Lin J, Scherlag BJ, Zhou J, Lu Z, Patterson E, Jackman WM, Lazzara R and Po SS.

Autonomic mechanism to explain complex fractionated atrial electrograms

(CFAE). J Cardiovasc Electrophysiol 2007;18:1197‐1205.

248

Lin YJ, Tai CT, Kao T, Tso HW, Higa S, Tsao HM, Chang SL, Hsieh MH and Chen SA.

Frequency analysis in different types of paroxysmal atrial fibrillation. J Am Coll

Cardiol 2006;47:1401‐1407.

Lin YJ, Tai CT, Kao T, Chang SL, Wongcharoen W, Lo LW, Tuan TC, Udyavar AR, Chen YJ,

Higa S, Ueng KC and Chen SA. Consistency of complex fractionated atrial

electrograms during atrial fibrillation. Heart Rhythm 2008;5:406‐412.

Littmann L, Svenson RH, Gallagher JJ, Bharati S, Lev M, Linder KD, Tatsis GP and

Nichelson C. Modification of sinus node function by epicardial laser irradiation

in dogs. Circulation 1990;81:350‐359.

Liu TY, Tai CT and Chen SA. Treatment of atrial fibrillation by catheter ablation of

conduction gaps in the crista terminalis and cavotricuspid isthmus of the right

atrium. J Cardiovasc Electrophysiol 2002;13:1044‐1046.

Lloyd‐Jones DM, Wang TJ, Leip EP, Larson MG, Levy D, Vasan RS, D'Agostino RB,

Massaro JM, Beiser A, Wolf PA and Benjamin EJ. Lifetime risk for development

of atrial fibrillation: the Framingham Heart Study. Circulation 2004;110:1042‐

1046.

Lo LW, Tai CT, Lin YJ, Chang SL, Wongcharoen W, Chang SH, Hsieh MH, Tuan TC,

Udyavar AR, Chen YJ, Tsao HM and Chen SA. Progressive remodeling of the

atrial substrate‐‐a novel finding from consecutive voltage mapping in patients

with recurrence of atrial fibrillation after catheter ablation. J Cardiovasc


Lu Z, Scherlag BJ, Lin J, Niu G, Ghias M, Jackman WM, Lazzara R, Jiang H and Po SS.

Autonomic mechanism for complex fractionated atrial electrograms: Evidence

by fast fourier transform analysis. J Cardiovasc Electrophysiol 2008;19:835‐842.

249

Luck JC and Engel TR. Dispersion of atrial refractoriness in patients with sinus node

dysfunction. Circulation 1979;60:404‐412.

Mackenzie J. The interpretation of the pulsations in the jugular veins. Am J Med Sci

1907;134:12.

Mandel W, Hayakawa H, Danzig R and Marcus HS. Evaluation of sino‐atrial node

function in man by overdrive suppression. Circulation 1971;44:59‐66.

Manios EG, Kanoupakis EM, Mavrakis HE, Kallergis EM, Dermitzaki DN and Vardas PE.

Sinus pacemaker function after cardioversion of chronic atrial fibrillation: is

sinus node remodeling related with recurrence? J Cardiovasc Electrophysiol

2001;12:800‐806.

Mansour M, Mandapati R, Berenfeld O, Chen J, Samie FH and Jalife J. Left‐to‐right

gradient of atrial frequencies during acute atrial fibrillation in the isolated

sheep heart. Circulation 2001;103:2631‐2636.

Marcus GM, Yang Y, Varosy PD, Ordovas K, Tseng ZH, Badhwar N, Lee BK, Lee RJ,

Scheinman MM and Olgin JE. Regional left atrial voltage in patients with atrial

fibrillation. Heart Rhythm 2007;4:138‐144.

Markides V, Schilling RJ, Ho SY, Chow AW, Davies DW and Peters NS. Characterization

of left atrial activation in the intact human heart. Circulation 2003;107:733‐739.

Markowitz SM, Nemirovksy D, Stein KM, Mittal S, Iwai S, Shah BK, Dobesh DP and

Lerman BB. Adenosine‐insensitive focal atrial tachycardia: evidence for de novo

micro‐re‐entry in the human atrium. J Am Coll Cardiol 2007;49:1324‐1333.

Marrouche NF, SippensGroenewegen A, Yang Y, Dibs S and Scheinman MM. Clinical

and electrophysiologic characteristics of left septal atrial tachycardia. J Am Coll

Cardiol 2002;40:1133‐1139.

250

Mary‐Rabine L, Albert A, Pham TD, Hordof A, Fenoglio JJ, Jr., Malm JR and Rosen MR.

The relationship of human atrial cellular electrophysiology to clinical function

and ultrastructure. Circ Res 1983;52:188‐199.

Matsuoka K, Kasai A, Fujii E, Omichi C, Okubo S, Teramura S, Uchida F and Nakano T.

Electrophysiological features of atrial tachycardia arising from the

atrioventricular annulus. Pacing Clin Electrophysiol 2002;25:440‐445.

McGuire MA, Bourke JP, Robotin MC, Johnson DC, Meldrum‐Hanna W, Nunn GR, Uther

JB and Ross DL. High resolution mapping of Koch's triangle using sixty

electrodes in humans with atrioventricular junctional (AV nodal) reentrant

tachycardia. Circulation 1993;88:2315‐2328.

McGuire MA, de Bakker JM, Vermeulen JT, Moorman AF, Loh P, Thibault B, Vermeulen

JL, Becker AE and Janse MJ. Atrioventricular junctional tissue. Discrepancy

between histological and electrophysiological characteristics. Circulation

1996;94:571‐577.

McWilliam JA. Fibrillar contraction of the heart. J Physiol 1886;8:296.

Meek WJ and Eyster JAE. Experiments on the origin and propagation of the impulse in

the heart. Heart 1914;5:227–244.

Meissner A, Christ M, Maagh P, Borchard R, van Bracht M, Wickenbrock I, Trappe HJ

and Plehn G. Quality of life and occurrence of atrial fibrillation in long‐term

follow‐up of common type atrial flutter ablation: ablation with irrigated 5 mm

tip and conventional 8 mm tip electrodes. Clin Res Cardiol 2007;96:794‐802.

Melnyk P, Ehrlich JR, Pourrier M, Villeneuve L, Cha TJ and Nattel S. Comparison of ion

channel distribution and expression in cardiomyocytes of canine pulmonary

veins versus left atrium. Cardiovasc Res 2005;65:104‐116.

251

Miyasaka Y, Barnes ME, Gersh BJ, Cha SS, Bailey KR, Abhayaratna WP, Seward JB and

Tsang TS. Secular Trends in Incidence of Atrial Fibrillation in Olmsted County,

Minnesota, 1980 to 2000, and Implications on the Projections for Future

Prevalence. Circulation 2006a.

Miyasaka Y, Barnes ME, Gersh BJ, Cha SS, Bailey KR, Abhayaratna W, Seward JB,

Iwasaka T and Tsang TS. Incidence and mortality risk of congestive heart failure

in atrial fibrillation patients: a community‐based study over two decades. Eur

Heart J 2006b;27:936‐941.

Miyasaka Y, Barnes ME, Petersen RC, Cha SS, Bailey KR, Gersh BJ, Casaclang‐Verzosa G,

Abhayaratna WP, Seward JB, Iwasaka T and Tsang TS. Risk of dementia in

stroke‐free patients diagnosed with atrial fibrillation: data from a community‐

based cohort. Eur Heart J 2007a;28:1962‐1967.

Miyasaka Y, Barnes ME, Bailey KR, Cha SS, Gersh BJ, Seward JB and Tsang TS. Mortality

trends in patients diagnosed with first atrial fibrillation: a 21‐year community‐

based study. J Am Coll Cardiol 2007b;49:986‐992.

Moe GK. On the multiple wavelet hypothesis of atrial fibrillation. Arch Int

Pharmacodyn Ther 1962;140:183‐188.

Moreira W, Timmermans C, Wellens HJ, Mizusawa Y, Philippens S, Perez D and

Rodriguez LM. Can common‐type atrial flutter be a sign of an arrhythmogenic

substrate in paroxysmal atrial fibrillation? Clinical and ablative consequences in

patients with coexistent paroxysmal atrial fibrillation/atrial flutter. Circulation

2007;116:2786‐2792.

Moreira W, Timmermans C, Wellens HJ, Mizusawa Y, Perez D, Philippens S and

Rodriguez LM. Long term outcome of cavotricuspid isthmus cryoablation for

252

the treatment of common atrial flutter in 180 patients: a single center

experience. J Interv Card Electrophysiol 2008;21:235‐240.

Morel E, Meyronet D, Thivolet‐Bejuy F and Chevalier P. Identification and distribution

of interstitial Cajal cells in human pulmonary veins. Heart Rhythm 2008;5:1063‐

1067.

Morillo CA, Klein GJ, Jones DL and Guiraudon CM. Chronic rapid atrial pacing.

Structural, functional, and electrophysiological characteristics of a new model

of sustained atrial fibrillation. Circulation 1995;91:1588‐1595.

Morton JB, Sanders P, Das A, Vohra JK, Sparks PB and Kalman JM. Focal atrial

tachycardia arising from the tricuspid annulus: electrophysiologic and

electrocardiographic characteristics. J Cardiovasc Electrophysiol 2001;12:653‐

659.

Morton JB, Byrne MJ, Power JM, Raman J and Kalman JM. Electrical remodeling of the

atrium in an anatomic model of atrial flutter: relationship between substrate

and triggers for conversion to atrial fibrillation. Circulation 2002;105:258‐264.

Morton JB, Sanders P, Vohra JK, Sparks PB, Morgan JG, Spence SJ, Grigg LE and Kalman

JM. Effect of chronic right atrial stretch on atrial electrical remodeling in

patients with an atrial septal defect. Circulation 2003;107:1775‐1782.

Moss AJ and Davis RJ. Brady‐Tachy syndrome. Prog Cardiovasc Dis 1974;16:439‐454.

Moubarak JB, Rozwadowski JV, Strzalka CT, Buck WR, Tan WS, Kish GF, Kisiel T, Fronc

HC and Maloney JD. Pulmonary veins‐left atrial junction: anatomic and

histological study. Pacing Clin Electrophysiol 2000;23:1836‐1838.

Murphy JG, Gersh BJ, McGoon MD, Mair DD, Porter CJ, Ilstrup DM, McGoon DC, Puga

FJ, Kirklin JW and Danielson GK. Long‐term outcome after surgical repair of

253

isolated atrial septal defect. Follow‐up at 27 to 32 years. N Engl J Med

1990;323:1645‐1650.

Nademanee K, McKenzie J, Kosar E, Schwab M, Sunsaneewitayakul B, Vasavakul T,

Khunnawat C and Ngarmukos T. A new approach for catheter ablation of atrial

fibrillation: mapping of the electrophysiologic substrate. J Am Coll Cardiol

2004;43:2044‐2053.

Nademanee K, Schwab MC, Kosar EM, Karwecki M, Moran MD, Visessook N, Michael

AD and Ngarmukos T. Clinical outcomes of catheter substrate ablation for high‐

risk patients with atrial fibrillation. J Am Coll Cardiol 2008;51:843‐849.

Nakagawa H, Lazzara R, Khastgir T, Beckman KJ, McClelland JH, Imai S, Pitha JV, Becker

AE, Arruda M, Gonzalez MD, Widman LE, Rome M, Neuhauser J, Wang X,

Calame JD, Goudeau MD and 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.

Nakagawa H, Shah N, Matsudaira K, Overholt E, Chandrasekaran K, Beckman KJ,

Spector P, Calame JD, Rao A, Hasdemir C, Otomo K, Wang Z, Lazzara R and

Jackman WM. Characterization of reentrant circuit in macroreentrant right

atrial tachycardia after surgical repair of congenital heart disease: isolated

channels between scars allow "focal" ablation. Circulation 2001;103:699‐709.

Narayan SM, Krummen DE, Kahn AM, Karasik PL and Franz MR. Evaluating fluctuations

in human atrial fibrillatory cycle length using monophasic action potentials.

Pacing Clin Electrophysiol 2006;29:1209‐1218.

254

Narula OS, Samet P and Javier RP. Significance of the sinus‐node recovery time.

Circulation 1972;45:140‐158.

Narula OS, Shantha N, Vasquez M, Towne WD and Linhart JW. A new method for

measurement of sinoatrial conduction time. Circulation 1978;58:706‐714.

Nathan H and Eliakim M. The junction between the left atrium and the pulmonary

veins. An anatomic study of human hearts. Circulation 1966;34:412‐422.

National Heart Foundation of Australia. Atrial Fibrillation. Information Sheet 305

(2004).

Ng J, Kadish AH and Goldberger JJ. Effect of electrogram characteristics on the

relationship of dominant frequency to atrial activation rate in atrial fibrillation.

Heart Rhythm 2006;3:1295‐1305.

Ng J, Kadish AH and Goldberger JJ. Technical considerations for dominant frequency

analysis. J Cardiovasc Electrophysiol 2007;18:757‐764.

O'Donnell D, Furniss SS and Bourke JP. Paroxysmal cycle length shortening in the

pulmonary veins during atrial fibrillation correlates with arrhythmogenic

triggering foci in sinus rhythm. J Cardiovasc Electrophysiol 2002;13:124‐128.

Off MK, Solheim E, Schuster P, Hoff PI, Ohm O‐J and Chen J. Detection of complex

fractionated electrograms by an automated algorithm (abstract). Heart Rhythm

2008;5:S171.

Okuyama Y, Miyauchi Y, Park AM, Hamabe A, Zhou S, Hayashi H, Miyauchi M, Omichi C,

Pak HN, Brodsky LA, Mandel WJ, Fishbein MC, Karagueuzian HS and Chen PS.

High resolution mapping of the pulmonary vein and the vein of Marshall during

induced atrial fibrillation and atrial tachycardia in a canine model of pacing‐

induced congestive heart failure. J Am Coll Cardiol 2003;42:348‐360.

255

Olgin JE, Kalman JM, Fitzpatrick AP and 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.

Olgin JE, Kalman JM and Lesh MD. Conduction barriers in human atrial flutter:

correlation of electrophysiology and anatomy. J Cardiovasc Electrophysiol

1996;7:1112‐1126.

Olgin JE, Kalman JM, Saxon LA, Lee RJ and Lesh MD. Mechanism of initiation of atrial

flutter in humans: site of unidirectional block and direction of rotation. J Am

Coll Cardiol 1997;29:376‐384.

Oral H, Ozaydin M, Tada H, Chugh A, Scharf C, Hassan S, Lai S, Greenstein R, Pelosi F,

Jr., Knight BP, Strickberger SA and Morady F. Mechanistic significance of

intermittent pulmonary vein tachycardia in patients with atrial fibrillation. J


Oral H, Chugh A, Good E, Wimmer A, Dey S, Gadeela N, Sankaran S, Crawford T,

Sarrazin JF, Kuhne M, Chalfoun N, Wells D, Frederick M, Fortino J, loucif‐Moore

S, Jongnarangsin K, Pelosi F, Jr., Bogun F and Morady F. Radiofrequency

catheter ablation of chronic atrial fibrillation guided by complex electrograms.

Circulation 2007;115:2606‐2612.

Ortiz J, Niwano S, Abe H, Rudy Y, Johnson NJ and Waldo AL. Mapping the conversion of

atrial flutter to atrial fibrillation and atrial fibrillation to atrial flutter. Insights

into mechanisms. Circulation Res 1994;74:882‐894.

256

Ott A, Breteler MM, de Bruyne MC, van Harskamp F, Grobbee DE and Hofman A. Atrial

fibrillation and dementia in a population‐based study. The Rotterdam Study.

Stroke 1997;28:316‐321.

Padeletti L, Pieragnoli P, Ciapetti C, Colella A, Musilli N, Porciani MC, Ricci R, Pignalberi

C, Santini M, Puglisi A, Azzolini P, Spampinato A, Martelli M, Capucci A, Boriani

G, Botto G and Proclemer A. Randomized crossover comparison of right atrial

appendage pacing versus interatrial septum pacing for prevention of

paroxysmal atrial fibrillation in patients with sinus bradycardia. Am Heart J

2001;142:1047‐1055.

Page PL, Plumb VJ, Okumura K and Waldo AL. A new animal model of atrial flutter. J

Am Coll Cardiol 1986;8:872‐879.

Papageorgiou P, Monahan K, Boyle NG, Seifert MJ, Beswick P, Zebede J, Epstein LM

and Josephson ME. Site‐dependent intra‐atrial conduction delay. Relationship

to initiation of atrial fibrillation. Circulation 1996;94:384‐389.

Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, Vicedomini G, Salvati A, Dicandia

C, Mazzone P, Santinelli V, Gulletta S and Chierchia S. Circumferential

radiofrequency ablation of pulmonary vein ostia: A new anatomic approach for

curing atrial fibrillation. Circulation 2000;102:2619‐2628.

Pappone C, Santinelli V, Manguso F, Vicedomini G, Gugliotta F, Augello G, Mazzone P,

Tortoriello V, Landoni G, Zangrillo A, Lang C, Tomita T, Mesas C, Mastella E and

Alfieri O. Pulmonary vein denervation enhances long‐term benefit after

circumferential ablation for paroxysmal atrial fibrillation. Circulation

2004;109:327‐334.

257

Paydak H, Kall JG, Burke MC, Rubenstein D, Kopp DE, Verdino RJ and Wilber DJ. Atrial

fibrillation after radiofrequency ablation of type I atrial flutter: time to onset,

determinants, and clinical course. Circulation 1998;98:315‐322.

Perez‐Lugones A, McMahon JT, Ratliff NB, Saliba WI, Schweikert RA, Marrouche NF,

Saad EB, Navia JL, McCarthy PM, Tchou P, Gillinov AM and Natale A. Evidence

of specialized conduction cells in human pulmonary veins of patients with atrial

fibrillation. J Cardiovasc Electrophysiol 2003;14:803‐809.

Peters NS, Cabo C and Wit AL (2000). Arrhythmogenic mechanisms: Automaticity,

Triggered Activity and Reentry. In: Zipes DP, Jalife J, editors. Cardiac

Electrophysiology: from Cell to Bedside: 345‐356.

Platonov PG. Interatrial conduction in the mechanisms of atrial fibrillation: from

anatomy to cardiac signals and new treatment modalities. Europace 2007;9

Suppl 6:vi10‐16.

Platonov PG, Mitrofanova L, Ivanov V and Ho SY. Substrates for intra‐atrial and

interatrial conduction in the atrial septum: anatomical study on 84 human

hearts. Heart Rhythm 2008;5:1189‐1195.

Porter M, Spear W, Akar JG, Helms R, Brysiewicz N, Santucci P and Wilber DJ.

Prospective study of atrial fibrillation termination during ablation guided by

automated detection of fractionated electrograms. J Cardiovasc Electrophysiol

2008;19:613‐620.

Power JM, Beacom GA, Alferness CA, Raman J, Wijffels M, Farish SJ, Burrell LM and

Tonkin AM. Susceptibility to atrial fibrillation: a study in an ovine model of

pacing‐induced early heart failure. J Cardiovasc Electrophysiol 1998;9:423‐435.

258

Prinzmetal MYRO, Corday ELIO, Brill IC, Sellers AL, Oblath RW, Flieg WA and Kruger HE.

Mechanism of the auricular arrhythmias. Circulation 1950;1:241‐245.

Psaty BM, Manolio TA, Kuller LH, Kronmal RA, Cushman M, Fried LP, White R, Furberg

CD and Rautaharju PM. Incidence of and risk factors for atrial fibrillation in

older adults. Circulation 1997;96:2455‐2461.

Ramanna H, de Bakker JM and Hauer RN. Mechanism of propensity to atrial fibrillation

in patients undergoing isthmus ablation for typical atrial flutter. J Cardiovasc


Ravelli F, Faes L, Sandrini L, Gaita F, Antolini R, Scaglione M and Nollo G. Wave

similarity mapping shows the spatiotemporal distribution of fibrillatory wave

complexity in the human right atrium during paroxysmal and chronic atrial


Rensma PL, Allessie MA, Lammers WJ, Bonke FI and Schalij MJ. Length of excitation

wave and susceptibility to reentrant atrial arrhythmias in normal conscious

dogs. Circulation Res 1988;62:395‐410.

Roberts‐Thomson KC, Stevenson IH, Kistler PM, Haqqani HM, Goldblatt JC, Sanders P

and Kalman JM. Anatomically determined functional conduction delay in the

posterior left atrium relationship to structural heart disease. J Am Coll Cardiol

2008;51:856‐862.

Roden DM, Lazzara R, Rosen M, Schwartz PJ, Towbin J and Vincent GM. Multiple

mechanisms in the long‐QT syndrome. Current knowledge, gaps, and future

directions. The SADS Foundation Task Force on LQTS. Circulation 1996;94:1996‐

2012.

259

Roithinger FX, Karch MR, Steiner PR, SippensGroenewegen A and Lesh MD.

Relationship between atrial fibrillation and typical atrial flutter in humans:

activation sequence changes during spontaneous conversion. Circulation

1997;96:3484‐3491.

Rosenbleuth A and Garcia‐Ramos J. Studies on flutter and fibrillation II. The influence

of artificial obstacles on experimental auricular flutter. Am Heart J 1947;33:677‐

684.

Rostagno C, Bacci F, Martelli M, Naldoni A, Bertini G and Gensini G. Clinical course of

lone atrial fibrillation since first symptomatic arrhythmic episode. Am J Cardiol

1995;76:837‐839.

Rostock T, Servatius H, Risius T, Ventura R, Weiss C, Meinertz T and Willems S. Impact

of amiodarone on electrophysiologic properties of pulmonary veins in patients

with paroxysmal atrial fibrillation. J Cardiovasc Electrophysiol 2005;16:39‐44.

Rostock T, Rotter M, Sanders P, Takahashi Y, Jaïs P, Hocini M, Hsu LF, Sacher F,

Clementy J and Haïssaguerre M. High‐density activation mapping of

fractionated electrograms in the atria of patients with paroxysmal atrial

fibrillation. Heart Rhythm 2006;3:27‐34.

Rostock T, Steven D, Lutomsky B, Servatius H, Drewitz I, Klemm H, Mullerleile K,

Ventura R, Meinertz T and Willems S. Atrial fibrillation begets atrial fibrillation

in the pulmonary veins on the impact of atrial fibrillation on the

electrophysiological properties of the pulmonary veins in humans. J Am Coll

Cardiol 2008;51:2153‐2160.

Rubenstein JJ, Schulman CL, Yurchak PM and DeSanctis RW. Clinical spectrum of the

sick sinus syndrome. Circulation 1972;46:5‐13.

260

Ryu K, Sahadevan J, Khrestian CM, Stambler BS and Waldo AL. Use of fast fourier

transform analysis of atrial electrograms for rapid characterization of atrial

activation‐implications for delineating possible mechanisms of atrial

tachyarrhythmias. J Cardiovasc Electrophysiol 2006;17:198‐206.

Saffitz JE, Kanter HL, Green KG, Tolley TK and Beyer EC. Tissue‐specific determinants of

anisotropic conduction velocity in canine atrial and ventricular myocardium.

Circ Res 1994;74:1065‐1070.

Saffitz JE, Green KG and Schuessler RB. Structural determinants of slow conduction in

the canine sinus node. J Cardiovasc Electrophysiol 1997;8:738‐744.

Sahadevan J, Ryu K, Peltz L, Khrestian CM, Stewart RW, Markowitz AH and Waldo AL.

Epicardial mapping of chronic atrial fibrillation in patients: preliminary

observations. Circulation 2004;110:3293‐3299.

Saito T, Waki K and Becker AE. Left atrial myocardial extension onto pulmonary veins in

humans: anatomic observations relevant for atrial arrhythmias. J Cardiovasc


Sanchez‐Quintana D, Cabrera JA, Farre J, Climent V, Anderson RH and Ho SY. Sinus

node revisited in the era of electroanatomical mapping and catheter ablation.

Heart 2005;91:189‐194.

Sanders P, Morton JB, Davidson NC, Spence SJ, Vohra JK, Sparks PB and Kalman JM.

Electrical remodeling of the atria in congestive heart failure:

electrophysiological and electroanatomic mapping in humans. Circulation

2003;108:1461‐1468.

Sanders P, Morton JB, Kistler PM, Spence SJ, Davidson NC, Hussin A, Vohra JK, Sparks

PB and Kalman JM. Electrophysiological and electroanatomic characterization

261

of the atria in sinus node disease: evidence of diffuse atrial remodeling.

Circulation 2004a;109:1514‐1522.

Sanders P, Kistler PM, Morton JB, Spence SJ and Kalman JM. Remodeling of sinus node

function in patients with congestive heart failure: reduction in sinus node

reserve. Circulation 2004b;110:897‐903.

Sanders P, Berenfeld O, Hocini M, Jaïs P, Vaidyanathan R, Hsu LF, Garrigue S, Takahashi

Y, Rotter M, Sacher F, Scavee C, Ploutz‐Snyder R, Jalife J and Haïssaguerre M.

Spectral analysis identifies sites of high‐frequency activity maintaining atrial

fibrillation in humans. Circulation 2005a;112:789‐797.

Sanders P, Hocini M, Jaïs P, Hsu LF, Takahashi Y, Rotter M, Scavee C, Pasquie JL, Sacher

F, Rostock T, Nalliah CJ, Clementy J and Haïssaguerre M. Characterization of

focal atrial tachycardia using high‐density mapping. J Am CollCardiol

2005b;46:2088‐2099.

Sanders P, Nalliah CJ, Dubois R, Takahashi Y, Hocini M, Rotter M, Rostock T, Sacher F,

Hsu LF, Jonsson A, O'Neill MD, Jaïs P and Haïssaguerre M. Frequency mapping

of the pulmonary veins in paroxysmal versus permanent atrial fibrillation. J


Saoudi N, Atallah G, Kirkorian G and Touboul P. Catheter ablation of the atrial

myocardium in human type I atrial flutter. Circulation 1990;81:762‐771.

Sarmast F, Kolli A, Zaitsev A, Parisian K, Dhamoon AS, Guha PK, Warren M, Anumonwo

JM, Taffet SM, Berenfeld O and Jalife J. Cholinergic atrial fibrillation: I(K,ACh)

gradients determine unequal left/right atrial frequencies and rotor dynamics.

Cardiovasc Res 2003;59:863‐873.

262

Scaglione M, Riccardi R, Calo L, Di DP, Lamberti F, Caponi D, Coda L and Gaita F. Typical

atrial flutter ablation: conduction across the posterior region of the inferior

vena cava orifice may mimic unidirectional isthmus block. J Cardiovasc


Scherf D. Studies on auricular tachycardia caused by aconitine administration. Proc Soc

Exp Biol Med 1947;64:233‐239.

Scherlag BJ, Yamanashi W, Patel U, Lazzara R and Jackman WM. Autonomically induced

conversion of pulmonary vein focal firing into atrial fibrillation. J Am Coll Cardiol

2005a;45:1878‐1886.

Scherlag BJ, Nakagawa H, Jackman WM, Yamanashi WS, Patterson E, Po S and Lazzara

R. Electrical stimulation to identify neural elements on the heart: their role in

atrial fibrillation. J IntervCard Electrophysiol 2005b;13 Suppl 1:37‐42.

Scherr D, Dalal D, Cheema A, Cheng A, Henrikson CA, Spragg D, Marine JE, Berger RD,

Calkins H and Dong J. Automated detection and characterization of complex

fractionated atrial electrograms in human left atrium during atrial fibrillation.

Heart Rhythm 2007;4:1013‐1020.

Schilling RJ, Peters NS and Davies DW. Simultaneous endocardial mapping in the

human left ventricle using a noncontact catheter: comparison of contact and

reconstructed electrograms during sinus rhythm. Circulation 1998;98:887‐898.

Schilling RJ, Kadish AH, Peters NS, Goldberger J and Davies DW. Endocardial mapping

of atrial fibrillation in the human right atrium using a non‐contact catheter. Eur

Heart J 2000;21:550‐564.

Schmieder S, Ndrepepa G, Dong J, Zrenner B, Schreieck J, Schneider MA, Karch MR and

Schmitt C. Acute and long‐term results of radiofrequency ablation of common

263

atrial flutter and the influence of the right atrial isthmus ablation on the

occurrence of atrial fibrillation. Eur Heart J 2003;24:956‐962.

Schmitt C, Estner H, Hecher B, Luik A, Kolb C, Karch M, Ndrepepa G, Zrenner B, Hessling

G and Deisenhofer I. Radiofrequency ablation of complex fractionated atrial

electrograms (CFAE): preferential sites of acute termination and regularization

in paroxysmal and persistent atrial fibrillation. J Cardiovasc Electrophysiol

2007;18:1039‐1046.

Schuessler RB, Boineau JP and Bromberg BI. Origin of the sinus impulse. J Cardiovasc


Schuessler RB, Kay MW, Melby SJ, Branham BH, Boineau JP and Damiano RJ, Jr. Spatial

and temporal stability of the dominant frequency of activation in human atrial

fibrillation. J Electrocardiol 2006;39:S7‐12.

Schumacher B, Jung W, Lewalter T, Vahlhaus C, Wolpert C and Luderitz B.

Radiofrequency ablation of atrial flutter due to administration of class IC

antiarrhythmic drugs for atrial fibrillation. Am J Cardiol 1999;83:710‐713.

Shah DC, Haïssaguerre M and Jaïs P. Toward a mechanism‐based understanding of

atrial fibrillation. J Cardiovasc Electrophysiol 2001;12:600‐601.

Shah DC, Sunthorn H, Burri H and Gentil‐Baron P. Evaluation of an individualized

strategy of cavotricuspid isthmus ablation as an adjunct to atrial fibrillation

ablation. J Cardiovasc Electrophysiol 2007;18:926‐930.

Shimizu A, Nozaki A, Rudy Y and Waldo AL. Onset of induced atrial flutter in the canine

pericarditis model. J Am Coll Cardiol 1991;17:1223‐1234.

264

Silverman ME. From rebellious palpitations to the discovery of auricular fibrillation:

contributions of Mackenzie, Lewis and Einthoven. Am J Cardiol 1994;73:384‐

389.

Skalidis EI, Hamilos MI, Karalis IK, Chlouverakis G, Kochiadakis GE and Vardas PE.

Isolated atrial microvascular dysfunction in patients with lone recurrent atrial

fibrillation. J Am Coll Cardiol 2008;51:2053‐2057.

Skanes AC, Mandapati R, Berenfeld O, Davidenko JM and Jalife J. Spatiotemporal

periodicity during atrial fibrillation in the isolated sheep heart. Circulation

1998;98:1236‐1248.

Soylu M, Demir AD, Ozdemir O, Topaloglu S, Aras D, Duru E, Sasmaz A and Korkmaz S.

Evaluation of atrial refractoriness immediately after percutaneous mitral

balloon commissurotomy in patients with mitral stenosis and sinus rhythm. Am

Heart J 2004;147:741‐745.

Spach MS, Miller WT, 3rd, Geselowitz DB, Barr RC, Kootsey JM and Johnson EA. The

discontinuous nature of propagation in normal canine cardiac muscle. Evidence

for recurrent discontinuities of intracellular resistance that affect the

membrane currents. Circ Res 1981;48:39‐54.

Spach MS and Dolber PC. 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. Circulation Res 1986;58:356‐371.

Sparks PB, Mond HG, Vohra JK, Jayaprakash S and Kalman JM. Electrical remodeling of

the atria following loss of atrioventricular synchrony: a long‐term study in


265

Sparks PB, Jayaprakash S, Vohra JK and Kalman JM. Electrical remodeling of the atria

associated with paroxysmal and chronic atrial flutter. Circulation

2000;102:1807‐1813.

Stevenson WG, Sager PT and Friedman PL. Entrainment techniques for mapping atrial

and ventricular tachycardias. J Cardiovasc Electrophysiol 1995;6:201‐216.

Stevenson WG and Stevenson LW. Atrial fibrillation in heart failure. N Engl J Med

1999;341:910‐911.

Stewart S, Hart CL, Hole DJ and McMurray JJ. Population prevalence, incidence, and

predictors of atrial fibrillation in the Renfrew/Paisley study. Heart 2001;86:516‐

521.

Stewart S, Hart CL, Hole DJ and McMurray JJ. A population‐based study of the long‐

term risks associated with atrial fibrillation: 20‐year follow‐up of the

Renfrew/Paisley study. Am J Med 2002;113:359‐364.

Stiles MK, Brooks AG, John B, Shashidhar, Wilson L, Kuklik P, Dimitri H, Lau DH,

Roberts‐Thomson RL, Mackenzie L, Willoughby S, Young GD and Sanders P. The

effect of electrogram duration on quantification of complex fractionated atrial

electrograms and dominant frequency. J Cardiovasc Electrophysiol

2008;19:252‐258.

Stiles MK, Brooks AG, Kuklik P, John B, Dimitri H, Lau DH, Wilson L, Shashidhar,

Roberts‐Thomson RL, Mackenzie L, Young GD and Sanders P. High‐density

mapping of atrial fibrillation in humans: relationship between high‐frequency

activation and electrogram fractionation. J Cardiovasc Electrophysiol online

early;doi:10.1111/j.1540‐8167.2008.01253.x.

266

Still AM, Raatikainen P, Ylitalo A, Kauma H, Ikaheimo M, Antero Kesaniemi Y and

Huikuri HV. Prevalence, characteristics and natural course of inappropriate

sinus tachycardia. Europace 2005;7:104‐112.

Strauss HC, Saroff AL, Bigger JT, Jr. and Giardina EG. Premature atrial stimulation as a

key to the understanding of sinoatrial conduction in man: presentation of data

and critical review of the literature. Circulation 1973;47:86‐93.

Tagawa M, Higuchi K, Chinushi M, Washizuka T, Ushiki T, Ishihara N and Aizawa Y.

Myocardium extending from the left atrium onto the pulmonary veins: a

comparison between subjects with and without atrial fibrillation. Pacing Clin


Tai CT, Chen SA, Chen YJ, Yu WC, Hsieh MH, Tsai CF, Chen CC, Ding YA and Chang MS.

Conduction properties of the crista terminalis in patients with typical atrial

flutter: basis for a line of block in the reentrant circuit. J Cardiovasc


Tai CT, Huang JL, Lee PC, Ding YA, Chang MS and Chen SA. High‐resolution mapping

around the crista terminalis during typical atrial flutter: new insights into

mechanisms. J Cardiovasc Electrophysiol 2004;15:406‐414.

Takahashi Y, Iesaka Y, Takahashi A, Goya M, Kobayashi K, Fujiwara H and Hiraoka M.

Reentrant tachycardia in pulmonary veins of patients with paroxysmal atrial


Takahashi Y, Jaïs P, Hocini M, Sanders P, Rotter M, Rostock T, Hsu LF, Sacher F,

Clementy J and Haïssaguerre M. Shortening of fibrillatory cycle length in the

pulmonary vein during vagal excitation. J Am CollCardiol 2006;47:774‐780.

267

Takahashi Y, O'Neill MD, Hocini M, Reant P, Jonsson A, Jaïs P, Sanders P, Rostock T,

Rotter M, Sacher F, Laffite S, Roudaut R, Clementy J and Haïssaguerre M.

Effects of stepwise ablation of chronic atrial fibrillation on atrial electrical and

mechanical properties. J Am Coll Cardiol 2007;49:1306‐1314.

Takahashi Y, O'Neill MD, Hocini M, Dubois R, Matsuo S, Knecht S, Mahapatra S, Lim KT,

Jaïs P, Jonsson A, Sacher F, Sanders P, Rostock T, Bordachar P, Clementy J, Klein

GJ and Haïssaguerre M. Characterization of electrograms associated with

termination of chronic atrial fibrillation by catheter ablation. J Am Coll Cardiol

2008;51:1003‐1010.

Thiagalingam A, Wallace EM, Boyd AC, Eipper VE, Campbell CR, Byth K, Ross DL and

Kovoor P. Noncontact mapping of the left ventricle: insights from validation

with transmural contact mapping. Pacing Clin Electrophysiol 2004a;27:570‐578.

Thiagalingam A, Wallace EM, Boyd AC, Eipper VE, Campbell CR, Byth K, Ross DL and

Kovoor P. Noncontact mapping of the left ventricle: insights from validation

with transmural contact mapping. Pacing Clin Electrophysiol 2004b;27:570‐578.

Tomita Y, Matsuo K, Sahadevan J, Khrestian CM and Waldo AL. Role of functional block

extension in lesion‐related atrial flutter. Circulation 2001;103:1025‐1030.

Tonkin AM and Heddle WF. Electrophysiological testing of sinus node function. Pacing

Clin Electrophysiol 1984;7:735‐748.

Triedman JK, Alexander ME, Berul CI, Bevilacqua LM and Walsh EP. Electroanatomic

mapping of entrained and exit zones in patients with repaired congenital heart

disease and intra‐atrial reentrant tachycardia. Circulation 2001;103:2060‐2065.

268

Tse HF, Lau CP and Ayers GM. Heterogeneous changes in electrophysiologic properties

in the paroxysmal and chronically fibrillating human atrium. J Cardiovasc


Verheule S, Wilson E, Everett T, Shanbhag S, Golden C and Olgin J. Alterations in atrial

electrophysiology and tissue structure in a canine model of chronic atrial

dilatation due to mitral regurgitation. Circulation 2003;107:2615‐2622.

Verma A, Wazni OM, Marrouche NF, Martin DO, Kilicaslan F, Minor S, Schweikert RA,

Saliba W, Cummings J, Burkhardt JD, Bhargava M, Belden WA, Abdul‐Karim A

and Natale A. Pre‐existent left atrial scarring in patients undergoing pulmonary

vein antrum isolation: an independent predictor of procedural failure. J Am Coll

Cardiol 2005;45:285‐292.

Verma A, Novak P, Macle L, Whaley B, Beardsall M, Wulffhart Z and Khaykin Y. A

prospective, multicenter evaluation of ablating complex fractionated

electrograms (CFEs) during atrial fibrillation (AF) identified by an automated

mapping algorithm: Acute effects on AF and efficacy as an adjuvant strategy.

Heart Rhythm 2008;5:198‐205.

Vidaillet H, Granada JF, Chyou PH, Maassen K, Ortiz M, Pulido JN, Sharma P, Smith PN

and Hayes J. A population‐based study of mortality among patients with atrial

fibrillation or flutter. Am J Med 2002;113:365‐370.

Volders PG, Zhu Q, Timmermans C, Eurlings PM, Su X, Arens YH, Li L, Jongbloed RJ, Xia

M, Rodriguez LM and Chen YH. Mapping a novel locus for familial atrial

fibrillation on chromosome 10p11‐q21. Heart Rhythm 2007;4:469‐475.

Volkmer M, Antz M, Hebe J and Kuck KH. Focal atrial tachycardia originating from the

musculature of the coronary sinus. J Cardiovasc Electrophysiol 2002;13:68‐71.

269

Waldo AI and Cooper TB. Spontaneous onset of type I atrial flutter in patients. J Am

Coll Cardiol 1996;28:707‐712.

Waldo AL, MacLean WA, Karp RB, Kouchoukos NT and James TN. Entrainment and

interruption of atrial flutter with atrial pacing: studies in man following open

heart surgery. Circulation 1977;56:737‐745.

Wang TJ, Larson MG, Levy D, Vasan RS, Leip EP, Wolf PA, D'Agostino RB, Murabito JM,

Kannel WB and Benjamin EJ. Temporal relations of atrial fibrillation and

congestive heart failure and their joint influence on mortality: the Framingham

Heart Study. Circulation 2003;107:2920‐2925.

Wazni O, Marrouche NF, Martin DO, Gillinov AM, Saliba W, Saad E, Klein A, Bhargava

M, Bash D, Schweikert R, Erciyes D, Abdul‐Karim A, Brachman J, Gunther J,

Pisano E, Potenza D, Fanelli R and Natale A. Randomized study comparing

combined pulmonary vein‐left atrial junction disconnection and cavotricuspid

isthmus ablation versus pulmonary vein‐left atrial junction disconnection alone

in patients presenting with typical atrial flutter and atrial fibrillation. Circulation

2003;108:2479‐2483.

Weiss C, Gocht A, Willems S, Hoffmann M, Risius T and Meinertz T. Impact of the

distribution and structure of myocardium in the pulmonary veins for

radiofrequency ablation of atrial fibrillation. Pacing Clin Electrophysiol

2002;25:1352‐1356.

Wijffels MC, Kirchhof CJ, Dorland R and Allessie MA. Atrial fibrillation begets atrial

fibrillation. A study in awake chronically instrumented goats. Circulation

1995;92:1954‐1968.

270

Willems S, Weiss C, Risius T, Rostock T, Hoffmann M, Ventura R and Meinertz T.

Dissociated activity and pulmonary vein fibrillation following functional

disconnection: impact for the arrhythmogenesis of focal atrial fibrillation.

Pacing Clin Electrophysiol 2003;26:1363‐1370.

Willems S and Rostock T. A basket full of mechanistic insights into pulmonary vein

arrhythmogenicity: reentry, focal activity, or fibrillatory conduction? J


Wolf PA, Dawber TR, Thomas HE, Jr. and Kannel WB. Epidemiologic assessment of

chronic atrial fibrillation and risk of stroke: the Framingham study. Neurology

1978;28:973‐977.

Wolf PA, Abbott RD and Kannel WB. Atrial fibrillation: a major contributor to stroke in

the elderly. The Framingham Study. Arch Intern Med 1987;147:1561‐1564.

Wolf PA, Abbott RD and Kannel WB. Atrial fibrillation as an independent risk factor for

stroke: the Framingham Study. Stroke 1991;22:983‐988.

Wolf PA, Benjamin EJ, Belanger AJ, Kannel WB, Levy D and D'Agostino RB. Secular

trends in the prevalence of atrial fibrillation: The Framingham Study. Am Heart

J 1996;131:790‐795.

Wu TJ, Doshi RN, Huang HL, Blanche C, Kass RM, Trento A, Cheng W, Karagueuzian HS,

Peter CT and Chen PS. Simultaneous biatrial computerized mapping during

permanent atrial fibrillation in patients with organic heart disease. J Cardiovasc


Wu TJ, Kerwin WF, Hwang C, Peter CT and Chen PS. Atrial fibrillation: focal activity, re‐

entry, or both? Heart Rhythm 2004;1:117‐120.

271

Yamane T, Shah DC, Hocini M, Jaïs P, Peng JT, Garrigue S, Clementy J and Haïssaguerre

M. Pulmonary vein aneurysm in association with arrhythmogenic foci in a

patient with focally initiated atrial fibrillation. J Cardiovasc Electrophysiol

2000;11:715.

Yamashita T, Inoue H, Nozaki A and Sugimoto T. Role of anatomic architecture in

sustained atrial reentry and double potentials. Am Heart J 1992;124:938‐946.

Yang Y, Mangat I, Glatter KA, Cheng J and Scheinman MM. Mechanism of conversion of

atypical right atrial flutter to atrial fibrillation. Am J Cardiol 2003;91:46‐52.

Yee R and Strauss HC. Electrophysiologic mechanisms: sinus node dysfunction.

Circulation 1987;75:III12‐18.

Yu WC, Lee SH, Tai CT, Tsai CF, Hsieh MH, Chen CC, Ding YA, Chang MS and Chen SA.

Reversal of atrial electrical remodeling following cardioversion of long‐standing

atrial fibrillation in man. Cardiovasc Res 1999;42:470‐476.

DVD ‘Movie 6.1’ is available in the University of Adelaide Library.

characterisation of the substrate of atrial fibrillation and flutter

Documents