Oxford Textbook of Interventional Cardiology [2 ed.] 9780198754152

Table of contents :
Contents
List of contributors
Section 1: Background and Basics
1 The epidemiology and pathophysiology of coronary artery disease • Robert Henderson and Richard Varcoe
2 The history of interventional cardiology • Toby Rogers, Kenneth Kent, and Augusto D. Pichard
3 Risk assessment and analysisof outcomes • Peter F. Ludman
4 Vascular access: femoral versus radial • Andrew Wiper and David H. Roberts
5 Radiation and percutaneous coronary intervention • Gurbir Bhatia and James Nolan
6 The ‘golden rules’ of percutaneous coronary intervention • Rod Stables
7 Care following percutaneous coronary intervention • Kevin O’Gallagher, Jonathan Byrne, and Philip MacCarthy
8 Trial design and interpretation in interventional cardiology: why is evidence-based medicine important? • Ayman Al-Saleh and Sanjit Jolly
Section 2: Percutaneous Coronary Intervention-related Imaging
9 Angiography: indications and limitations • David Adlam, Annette Maznyczka, and Bernard Prendergast
10 Coronary physiology in clinical practice • Olivier Muller, Stephane Fournier, and Bernard De Bruyne
11 The role of intravascular ultrasound in percutaneous coronary intervention • Kozo Okada, Yasuhiro Honda, and Peter J. Fitzgerald
12 Intravascular ultrasound and optical coherence tomography in percutaneous coronary intervention • Ravinay Bhindi, Usaid K. Allahwala, and Keith M. Channon
13 Cardiac computed tomography for the interventionalist • Adriaan Coenen, Laurens E. Swart, Ricardo P. J. Budde, and Koen Nieman
14 Cardiovascular magnetic resonance • Theodoros D. Karamitsos and Stefan Neubauer
Section 3: Percutaneous Coronary Intervention by Clinical Syndrome
15 Stable coronary disease: medical therapy versus percutaneous coronary intervention versus surgery • Vasim Farooq and Patrick W. Serruys
16 Percutaneous coronary intervention in non-ST elevation acute coronary syndrome • Bashir Alaour, Michael Mahmoudi, and Nick Curzen
17 Primary percutaneous coronary intervention for ST-elevation myocardial infarction • Zulfiquar Adam and Mark A. de Belder
18 Percutaneous coronary intervention in patients with impaired left ventricular function • Divaka Perera and Natalia Briceno
Section 4: Percutaneous Coronary Intervention by Lesion and Patient Subsets
19 Coronary bifurcationstenting: state of the art • Yves Louvard, Philippe Garot, Thomas Hovasse, Bernard Chevalier, and Thierry Lefèvre
20 Percutaneous coronary intervention for left main coronary artery stenosis • Michael Mahmoudi, Nick Curzen, Christine Hughes, Bruno Farah, and Jean Fajadet
21 Chronic total occlusions • Colm G. Hanratty, James C. Spratt, and Simon J. Walsh
22 Revascularization in patients with diabetes mellitus • George Kassimis and Adrian Banning
23 Out-of-hospital cardiac arrest: role of percutaneous coronary intervention • Peter Radsel and Marko Noc
Section 5: Adjunctive Drug Therapies in Percutaneous Coronary Intervention
24 Oral antiplatelet therapies in percutaneous coronary intervention • Vikram Khanna, Tony Gershlick, and Nick Curzen
25 Current status of glycoprotein IIb/IIIa inhibitors • Tim Lockie
26 The role of bivalirudin in percutaneous coronary intervention • Stefanie Schüpke, Steffen Massberg, and Adnan Kastrati
27 Optimal medical therapy in percutaneous coronary intervention patients: statins and angiotensin-converting enzyme inhibitors as disease-modifying agents • Simon J. Corbett and Nick Curzen
28 New oral anticoagulants: the evidence base and role in patients undergoing percutaneous coronary intervention • Mikhail S. Dzeshka, Richard A. Brown, and Gregory Y. H. Lip
Section 6: Complications of Percutaneous Coronary Intervention
29 Contrast-induced acute kidney injury • Peter A. McCullough
30 In-stent restenosis in the drug-eluting stent era • Jaya Chandrasekhar, Adriano Caixeta, Philippe Généreux, George Dangas, and Roxana Mehran
31 Stent thrombosis • Nikesh Malik, Amerjeet Banning, and Tony Gershlick
32 Stent loss and retrieval • Mrinal Saha and Adam de Belder
33 Coronary artery perforation • Mark Gunning and Chee Wah Khoo
Section 7: Special Devices in Percutaneous Coronary Intervention
34 Rotational atherectomy • Adam de Belder, Martyn Thomas, and Emanuele Barbato
35 Laser • Peter O’Kane and Simon Redwood
36 No-reflow in native coronaries and vein grafts: the role of drugs and distal protection • Giovanni Luigi De Maria and Adrian Banning
37 Bioresorbable vascular scaffolds • Adam J. Brown and Nick E. J. West
38 Access routes and the transcatheter aortic valve implantation procedure • Corrado Tamburino, Claudia Ina Tamburino, and Sebastiano Immè
39 Selection of transcatheter aortic valve implantation prostheses • Mohamed Abdel- Wahab and John Jose
40 Transcatheter aortic valve implantation and the management of coronary artery disease • Muhammed Zeeshan Khawaja and Simon Redwood
41 Transcatheter mitral valve repair • Michael Bellamy and Christopher Baker
42 Transcatheter mitral valve replacement • Ricardo Boix Garibo, Mohsin Uzzaman, Michael Ghosh- Dastidar, and Vinayak Bapat
Section 8: Non-coronary Percutaneous Interventions
43 Percutaneous device closure of atrial septal defect and patent foramen ovale • Patrick A. Calvert, Bushra S. Rana, Roland Hilling-Smith, and David Hildick-Smith
44 Mitral balloon valvuloplasty • Alec Vahanian, Dominique Himbert, Eric Brochet, Grégory Ducrocq, and Bernard Iung
45 Alcohol septal ablation for obstructive hypertrophic cardiomyopathy • Charles Knight, Saidi Mohiddin, and Constantinos O’Mahony
46 Carotid artery stenting • Iqbal Malik and Mohamed Hamady
47 Left atrial appendage occlusion • Sandeep Panikker, Tim Betts, and Milena Leo
Section 9: The Future
48 Novel device therapies for resistant hypertension • Kenneth Chan, Manish Saxena, and Melvin D. Lobo
49 Robotic percutaneous coronary intervention • Giora Weisz
50 Stem cell delivery and therapy • Fizzah Choudry and Anthony Mathur
Index

Citation preview

i

Oxford Textbook of

Interventional Cardiology

ii

iii

Oxford Textbook of

Interventional Cardiology SECOND EDITION

Edited by

Simon Redwood Nick Curzen and

Adrian Banning

1

iv

1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Oxford University Press 2018 The moral rights of the authors‌have been asserted First Edition published in 2010 Second Edition published in 2018 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2018943513 ISBN 978–​0–​19–​875415–​2 Printed in Great Britain by Bell & Bain Ltd., Glasgow Oxford University Press makes no representation, express or implied, that the drug dosages in this book are correct. Readers must therefore always check the product information and clinical procedures with the most up-​to-​date published product information and data sheets provided by the manufacturers and the most recent codes of conduct and safety regulations. The authors and the publishers do not accept responsibility or legal liability for any errors in the text or for the misuse or misapplication of material in this work. Except where otherwise stated, drug dosages and recommendations are for the non-​pregnant adult who is not breast-​feeding Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

v

Contents

List of contributors  ix

SECTION 2

SECTION 1

Percutaneous Coronary Intervention-​ related Imaging  

Background and Basics   1 The epidemiology and pathophysiology of coronary artery disease   3 Robert Henderson and Richard Varcoe

2 The history of interventional cardiology   15 Toby Rogers, Kenneth Kent, and Augusto D. Pichard

3 Risk assessment and analysis of outcomes   25 Peter F. Ludman

4 Vascular access: femoral versus radial   49 Andrew Wiper and David H. Roberts

5 Radiation and percutaneous coronary intervention   65 Gurbir Bhatia and James Nolan

6 The ‘golden rules’ of percutaneous coronary intervention   75 Rod Stables

7 Care following percutaneous coronary intervention   81 Kevin O’Gallagher, Jonathan Byrne, and Philip MacCarthy

8 Trial design and interpretation in interventional cardiology: why is evidence-​based medicine important?   91 Ayman Al-​Saleh and Sanjit Jolly

9 Angiography: indications and limitations   99 David Adlam, Annette Maznyczka, and Bernard Prendergast

10 Coronary physiology in clinical practice   127 Olivier Muller, Stephane Fournier, and Bernard De Bruyne

11 The role of intravascular ultrasound in percutaneous coronary intervention   145 Kozo Okada, Yasuhiro Honda, and Peter J. Fitzgerald

12 Intravascular ultrasound and optical coherence tomography in percutaneous coronary intervention   171 Ravinay Bhindi, Usaid K. Allahwala, and Keith M. Channon

13 Cardiac computed tomography for the interventionalist   177 Adriaan Coenen, Laurens E. Swart, Ricardo P. J. Budde, and Koen Nieman

14 Cardiovascular magnetic resonance   191 Theodoros D. Karamitsos and Stefan Neubauer

SECTION 3

Percutaneous Coronary Intervention by Clinical Syndrome   15 Stable coronary disease: medical therapy versus percutaneous coronary intervention versus surgery   211 Vasim Farooq and Patrick W. Serruys

vi

vi

contents

16 Percutaneous coronary intervention in non-​ST elevation acute coronary syndrome   235 Bashir Alaour, Michael Mahmoudi, and Nick Curzen

17 Primary percutaneous coronary intervention for ST-​elevation myocardial infarction   251 Zulfiquar Adam and Mark A. de Belder

18 Percutaneous coronary intervention in patients with impaired left ventricular function   273

27 Optimal medical therapy in percutaneous coronary intervention patients: statins and angiotensin-​converting enzyme inhibitors as disease-​modifying agents   417 Simon J. Corbett and Nick Curzen

28 New oral anticoagulants: the evidence base and role in patients undergoing percutaneous coronary intervention   433

Divaka Perera and Natalia Briceno

Mikhail S. Dzeshka, Richard A. Brown, and Gregory Y. H. Lip

SECTION 4

SECTION 6

Percutaneous Coronary Intervention by Lesion and Patient Subsets   19 Coronary bifurcation stenting: state of the art   287 Yves Louvard, Philippe Garot, Thomas Hovasse, Bernard Chevalier, and Thierry Lefèvre

20 Percutaneous coronary intervention for left main coronary artery stenosis   305 Michael Mahmoudi, Nick Curzen, Christine Hughes, Bruno Farah, and Jean Fajadet

21 Chronic total occlusions   315 Colm G. Hanratty, James C. Spratt, and Simon J. Walsh

22 Revascularization in patients with diabetes mellitus   337 George Kassimis and Adrian Banning

23 Out-​of-​hospital cardiac arrest: role of percutaneous coronary intervention   351 Peter Radsel and Marko Noc

SECTION 5

Adjunctive Drug Therapies in Percutaneous Coronary Intervention   24 Oral antiplatelet therapies in percutaneous coronary intervention   363 Vikram Khanna, Tony Gershlick, and Nick Curzen

25 Current status of glycoprotein IIb/​IIIa inhibitors   381 Tim Lockie

26 The role of bivalirudin in percutaneous coronary intervention   401 Stefanie Schüpke, Steffen Massberg, and Adnan Kastrati

Complications of Percutaneous Coronary Intervention   29 Contrast-​induced acute kidney injury   447 Peter A. McCullough

30 In-​stent restenosis in the drug-​eluting stent era   453 Jaya Chandrasekhar, Adriano Caixeta, Philippe Généreux, George Dangas, and Roxana Mehran

31 Stent thrombosis   473 Nikesh Malik, Amerjeet Banning, and Tony Gershlick

32 Stent loss and retrieval   501 Mrinal Saha and Adam de Belder

33 Coronary artery perforation   511 Mark Gunning and Chee Wah Khoo

SECTION 7

Special Devices in Percutaneous Coronary Intervention   34 Rotational atherectomy   525 Adam de Belder, Martyn Thomas, and Emanuele Barbato

35 Laser   535 Peter O’Kane and Simon Redwood

36 No-​reflow in native coronaries and vein grafts: the role of drugs and distal protection   557 Giovanni Luigi De Maria and Adrian Banning

37 Bioresorbable vascular scaffolds   569 Adam J. Brown and Nick E. J. West

38 Access routes and the transcatheter aortic valve implantation procedure   583 Corrado Tamburino, Claudia Ina Tamburino, and Sebastiano Immè

vi

 

39 Selection of transcatheter aortic valve implantation prostheses   589 Mohamed Abdel-​Wahab and John Jose

40 Transcatheter aortic valve implantation and the management of coronary artery disease   601 Muhammed Zeeshan Khawaja and Simon Redwood

41 Transcatheter mitral valve repair   607 Michael Bellamy and Christopher Baker

42 Transcatheter mitral valve replacement   629 Ricardo Boix Garibo, Mohsin Uzzaman, Michael Ghosh-​Dastidar, and Vinayak Bapat

SECTION 8

Non-​coronary Percutaneous Interventions   43 Percutaneous device closure of atrial septal defect and patent foramen ovale   641 Patrick A. Calvert, Bushra S. Rana, Roland Hilling-​Smith, and David Hildick-​Smith

44 Mitral balloon valvuloplasty   657 Alec Vahanian, Dominique Himbert, Eric Brochet, Grégory Ducrocq, and Bernard Iung

contents

45 Alcohol septal ablation for obstructive hypertrophic cardiomyopathy   669 Charles Knight, Saidi Mohiddin, and Constantinos O’Mahony

46 Carotid artery stenting   681 Iqbal Malik and Mohamed Hamady

47 Left atrial appendage occlusion   703 Sandeep Panikker, Tim Betts, and Milena Leo

SECTION 9

The Future   48 Novel device therapies for resistant hypertension   717 Kenneth Chan, Manish Saxena, and Melvin D. Lobo

49 Robotic percutaneous coronary intervention   731 Giora Weisz

50 Stem cell delivery and therapy   737 Fizzah Choudry and Anthony Mathur

Index   751

vii

vi

ix

List of contributors

Mohamed Abdel-​Wahab Segeberger Kliniken GmbH (Academic Teaching Hospital of the Universities of Kiel, Lübeck and Hamburg), Bad Segeberg, Germany Zulfiquar Adam The James Cook University Hospital, Middlesbrough, UK David Adlam Cardiovascular Research Centre, University of Leicester, Leicester, UK Bashir Alaour Faculty of Medicine, University of Southampton, Southampton, UK

Richard A. Brown Institute of Cardiovascular Sciences, University of Birmingham, Birmingham, UK Ricardo P. J. Budde Erasmus University Medical Center, Rotterdam, the Netherlands Jonathan Byrne King’s College Hospital, London, UK Adriano Caixeta Hospital Israelita Albert Einstein, São Paulo, Brazil Patrick A. Calvert Royal Papworth Hospital, Cambridge, UK

Usaid K. Allahwala Northern Clinical School, University of Sydney, Australia

Kenneth Chan University College London, Royal Free Hospital, London, UK

Ayman Al-​Saleh Department of Medicine, McMaster University, Ontario, Canada; Department of Cardiology, King Saud university, Saudi Arabia

Jaya Chandrasekhar Ichahn School of Medicine, Mount Sinai, New York, USA

Christopher Baker Imperial Healthcare NHS Trust, Hammersmith Hospital, London, UK

Keith M. Channon Radcliffe Department of Medicine, University of Oxford, Oxford, UK

Adrian Banning John Radcliffe Hospital, Oxford, UK

Bernard Chevalier Institut Cardiovasculaire Paris Sud, Massy, France

Amerjeet Banning St George’s Hospital Medical School, London, UK

Fizzah Choudry Barts Heart Centre, St Bartholomew’s Hospital, London, UK

Vinayak Bapat St Thomas’ Hospital, London, UK

Adriaan Coenen Erasmus Medical Center, Rotterdam, the Netherlands

Emanuele Barbato Cardiovascular Center Aalst, Aalst, Belgium Michael Bellamy Imperial Healthcare NHS Trust, Hammersmith Hospital, London, UK Tim Betts John Radcliffe Hospital, Oxford, UK Gurbir Bhatia Birmingham Heartlands Hospital, UK Ravinay Bhindi Northern Clinical School, University of Sydney, Australia Ricardo Boix Garibo St Thomas’ Hospital, London, UK Natalia Briceno Cardiovascular Clinical Academic Group, St Thomas’ Hospital, London, UK Eric Brochet Cardiology Department, Bichat Hospital, University Paris VII, Paris, France Adam J. Brown Monash Cardiovascular Research Centre, Monash University & MonashHeart, Melbourne, Australia

Simon J. Corbett University Hospital Southampton, University of Southampton, Southampton, UK Nick Curzen Faculty of Medicine, University of Southampton, Southampton, UK George Dangas Zena and Michael A. Weiner Cardiovascular Institute, Mount Sinai School of Medicine, New York, USA Adam de Belder Brighton and Sussex University Hospitals NHS Trust, UK Mark A. de Belder The James Cook University Hospital, Middlesbrough, UK Bernard De Bruyne Cardiovascular Center Aalst, Aalst, Belgium Giovanni Luigi De Maria John Radcliffe Hospital, Oxford, UK Grégory Ducrocq Cardiology Department, Bichat Hospital, University Paris VII, Paris, France

x

x

list of contributors Mikhail S. Dzeshka Institute of Cardiovascular Sciences, University of Birmingham, Birmingham, UK, and Grodno State Medical University, Grodno, Belarus Jean Fajadet Clinique Pasteur, Toulouse, France Bruno Farah Clinique Pasteur, Toulouse, France Vasim Farooq St George’s Hospital, London, UK Peter J. Fitzgerald Division of Cardiovascular Medicine, Stanford University School of Medicine, California, USA Stephane Fournier Department of Cardiology, University Hospital, Lausanne, Switzerland

Adnan Kastrati Deutsches Herzzentrum München, Technische Universität, Munich, Germany Kenneth Kent Medstar Heart Institute, Washington, District of Columbia, USA Vikram Khanna Faculty of Medicine, University of Southampton, Southampton, UK Charles Knight Barts Heart Centre, St Bartholomew’s Hospital, London, UK Thierry Lefèvre Institut Cardiovasculaire Paris Sud, Massy, France

Philippe Généreux Morristown Medical Centre, New Jersey, USA

Gregory Y. H. Lip Institute of Cardiovascular Sciences, University of Birmingham, Birmingham, UK, and Aalborg Thrombosis Research Unit, Department of Clinical Medicine, Faculty of Health, Aalborg University, Aalborg, Denmark

Tony Gershlick Department of Cardiovascular Sciences, University of Leicester, Leicester, UK

Milena Leo Cardiac Electrophysiology Research Fellow, Oxford University Hospitals NHS Foundation Trust, Oxford, UK

Michael Ghosh-​Dastidar St Thomas’ Hospital, London, UK

Melvin D. Lobo William Harvey Research Institute, Queen Mary University of London, London, UK

Philippe Garot Institut Cardiovasculaire Paris Sud, Massy, France

Mark Gunning Royal Stoke University Hospital, UK Mohamed Hamady Imperial Healthcare NHS Trust, Hammersmith Hospital, London, UK

Tim Lockie Royal Free Hospital, London, UK Yves Louvard Institut Cardiovasculaire Paris Sud, Massy, France

Colm G. Hanratty Belfast Health and Social Care Trust, Belfast, UK

Peter F. Ludman Queen Elizabeth Hospital Birmingham, Birmingham, UK

Robert Henderson Nottingham University Hospitals, Nottingham, UK

Philip MacCarthy King’s College Hospital, London, UK

David Hildick-​Smith Royal Sussex County Hospital, Brighton, UK

Michael Mahmoudi Faculty of Medicine, University of Southampton, Southampton, UK

Roland Hilling-​Smith Queensland Cardiovascular Group, Mater Hospital, Brisbane, Australia

Iqbal Malik Imperial Healthcare NHS Trust, Hammersmith Hospital, London, UK

Dominique Himbert Cardiology Department, Bichat Hospital, University Paris VII, Paris, France

Nikesh Malik University of Leicester, Leicester, UK

Yasuhiro Honda Division of Cardiovascular Medicine, Stanford University School of Medicine, California, USA Thomas Hovasse Institut Cardiovasculaire Paris Sud, Massy, France Christine Hughes Clinique Pasteur, Toulouse, France Sebastiano Immè Ferrarotto Hospital, University of Catania, Catania, Italy Bernard Iung Cardiology Department, Bichat Hospital, University Paris VII, Paris, France Sanjit Jolly Department of Medicine, McMaster University, Ontario, Canada John Jose Segeberger Kliniken GmbH (Academic Teaching Hospital of the Universities of Kiel, Lübeck and Hamburg), Bad Segeberg, Germany Theodoros D. Karamitsos Aristotle University of Thessaloniki, Thessaloniki, Greece George Kassimis Cheltenham General Hospital, Cheltenham, UK

Steffen Massberg Ludwig-​Maximilians-​Universität München, Munich, Germany Anthony Mathur Barts Heart Centre, St Bartholomew’s Hospital, London, UK Annette Maznyczka King’s College Hospital, London, UK Peter A. McCullough Baylor University Medical Center, Dallas, USA Roxana Mehran Zena and Michael A. Weiner Cardiovascular Institute, Mount Sinai School of Medicine, New York, USA Saidi Mohiddin Barts Heart Centre, St Bartholomew’s Hospital, London, UK Olivier Muller Department of Cardiology, University Hospital, Lausanne, Switzerland Stefan Neubauer Oxford Centre for Clinical Magnetic Resonance Research, John Radcliffe Hospital, Oxford, UK Koen Nieman Erasmus Medical Center, Rotterdam, the Netherlands, and Stanford University School of Medicine, Stanford, USA

xi

 

list of contributors

Marko Noc Center for Intensive Internal Medicine, University Medical Center, Ljubljana, Slovenia

Stefanie Schüpke Deutsches Herzzentrum München, Technische Universität, Munich, Germany

James Nolan Royal Stoke University Hospital, Stoke-​on-​Trent, UK

Patrick W. Serruys Thoraxcenter, Erasmus MC, Rotterdam, The Netherlands

Kevin O’Gallagher King’s College Hospital, London, UK Peter O’Kane Royal Bournemouth Hospital, Bournemouth, UK Constantinos O’Mahony Barts Heart Centre, St Bartholomew’s Hospital, London, UK Kozo Okada Division of Cardiovascular Medicine, Stanford University School of Medicine, California, USA Sandeep Panikker Royal Brompton Hospital, London, UK Divaka Perera Cardiovascular Clinical Academic Group, St Thomas’ Hospital, London, UK Augusto D. Pichard Medstar Heart Institute, Washington, District of Columbia, USA Bernard Prendergast St Thomas’ Hospital, London, UK Peter Radsel Center for Intensive Internal Medicine, University Medical Center, Ljubljana, Slovenia

James C. Spratt Belfast Health and Social Care Trust, Belfast, UK Rod Stables Liverpool Heart and Chest Hospital, Liverpool, UK Laurens E. Swart Erasmus Medical Center, Rotterdam, the Netherlands Claudia Ina Tamburino Ferrarotto Hospital, University of Catania, Catania, Italy Corrado Tamburino Ferrarotto Hospital, University of Catania, Catania, Italy Martyn Thomas St Thomas’ Hospital, London, UK Mohsin Uzzaman Birmingham Children’s Hospital, Birmingham, UK Alec Vahanian Cardiology Department, Bichat Hospital, University Paris VII, Paris, France

Bushra S. Rana Royal Papworth Hospital, Cambridge, UK

Richard Varcoe Nottingham University Hospitals, Nottingham, UK

Simon Redwood King’s College London, St Thomas’ Hospital, London, UK

Chee Wah Khoo Royal Stoke University Hospital, UK

David H. Roberts Lancashire Cardiac Centre, Blackpool Victoria Hospital, Blackpool, UK Toby Rogers Medstar Heart Institute, Washington, District of Columbia, USA Mrinal Saha Consultant Cardiologist, Cheltenham and Gloucester Hospitals NHS Trust, UK Manish Saxena Barts Heart Centre, Barts Health NHS Trust, London, UK

Simon J. Walsh Belfast Health and Social Care Trust, Belfast, UK Giora Weisz Montefiore-​Einstein Center for Heart and Vascular Care, New York, USA Nick E. J. West Royal Papworth Hospital, Cambridge, UK Andrew Wiper Lancashire Cardiac Centre, Blackpool, UK Muhammed Zeeshan Khawaja Guy’s and Thomas’ NHS Hospitals Foundation Trust, London, UK

xi

xi

1

SECTION 1

Background and Basics

1 The epidemiology and pathophysiology of coronary artery disease   3 Robert Henderson and Richard Varcoe 2 The history of interventional cardiology   15 Toby Rogers, Kenneth Kent, and Augusto D. Pichard 3 Risk assessment and analysis of outcomes   25 Peter F. Ludman 4 Vascular access: femoral versus radial   49 Andrew Wiper and David H. Roberts 5 Radiation and percutaneous coronary intervention   65 Gurbir Bhatia and James Nolan

6 The ‘golden rules’ of percutaneous coronary intervention   75 Rod Stables 7 Care following percutaneous coronary intervention   81 Kevin O’Gallagher, Jonathan Byrne, and Philip MacCarthy 8 Trial design and interpretation in interventional cardiology: why is evidence-​based medicine important?   91 Ayman Al-​Saleh and Sanjit Jolly

2

3

CHAPTER 1

The epidemiology and pathophysiology of coronary artery disease Robert Henderson and Richard Varcoe

Epidemiology of coronary heart disease Advances in the prevention and treatment of ischaemic heart disease (IHD) have led to significant improvements in prognosis and quality of life. However, the ageing and growth of populations has led to an increase in the total number of deaths, and IHD remains a leading global cause of premature death and disability. From 1990 to 2013 age-​standardized global death rates from IHD fell from 177·3 to 137·8 per 100,000, but the total number of deaths due to IHD increased from 5.7  million to 8.1  million. Coronary heart disease (CHD) has risen from fourth to first in the rank of causes of global years of life lost (1) and is projected to remain as the leading cause of death, accounting for 13.4% of all deaths in 2030 (2).

Mortality In the UK age-​standardized death rates from CHD have declined over several decades. From 1974 to 2013, age-​standardized CHD death rates declined by 73% in those dying at any age and 81% for those dying before age 75. Nevertheless, CHD remains the biggest single cause of death in the UK, accounting for around one in seven deaths in men and one in ten deaths in women, and in 2014 was responsible for around 22,300 deaths under the age of 75 years (3). In the USA annual CHD mortality declined by 39.2% from 2000 to 2010 (the actual number of deaths fell by only 26.3%), but in 2010 CHD accounted for one in six of all deaths (4). In high-​income countries there are substantial regional, social, and ethnic variations in coronary disease-​associated mortality. For example, in the UK in 2011–​13 the age-​standardized CHD death rate in Scotland was 45% higher overall and 72% higher for premature deaths than the rates for south-​east England. Death rates from CHD increase during the winter months, and in 2012–​13 the winter CHD mortality in England was 19% higher than at other times of the year (3). In recent years the decline in coronary mortality in high-​income countries has been slower in younger than in older age groups. For example, in the UK from 1997 to 2006 there was a 46% fall in CHD mortality amongst men aged 55–​64 years but only a 22% fall amongst men aged 35–​44 years (5). In the USA the decline in age-​ adjusted coronary mortality from 1980 to 2002 slowed markedly

in adults aged 35–​54 years. Moreover, since 1997 the mortality rate among women aged 35–​44 has been increasing by about 1.3% per year (6). The decline in the rate of death from cardiovascular disease in several high-​income countries has been attributed to reductions in risk factors and improved management of cardiovascular disease (7). It has been estimated that 58% of the decline in coronary mortality in the UK between 1981 and 2000 was attributable to reductions in major risk factors, principally smoking, but the remaining 42% was explained by treatment of individuals, including secondary prevention (8). In the USA 47% of the reduction in CHD mortality from 1980 to 2000 has been attributed to treatments and 44% to modification of risk factors, but these reductions were partially offset by a rise in mortality attributable to increases in body mass index and diabetes prevalence (9). The World Health Organization (WHO) MONICA project examined temporal trends in cardiovascular mortality over the 1980s and 1990s in 21 countries, and demonstrated a strong link between improved care for patients with myocardial infarction and the decline in coronary mortality (10). An investigation into the potential impact of various preventative and interventional strategies on CHD-​related mortality in the USA estimated that delivery of ‘perfect care’ (through the modification of risk factors and use of all effective therapies) to a hypothetical population (aged 30–​84 years) could prevent or postpone around 75% of cardiac deaths (11). Globally, IHD mortality rates vary more than 20-​fold between countries. In high-​income countries age-​standardized IHD mortality has been steadily declining over several decades, but population growth and ageing have maintained a high disease burden. By contrast, in some low-​and middle-​income countries IHD mortality rates are stable or are increasing, especially amongst younger adults adopting urbanized lifestyles. The highest IHD mortality rates are reported in Eastern Europe and Central Asia, and low-​ and middle-​income countries now account for more than 80% of global IHD deaths (12–​14) (Fig 1.1). Age-​standardized IHD disability-​adjusted life years (DALYs—​ years of life lost to premature deaths and years lived with non-​fatal disease or disability) decreased in most countries between 1990 and 2010, but increased in several countries in Eastern Europe and Central and South Asia. In 2010 around two-​thirds of IHD DALYs

4

4

Section 1  

background and basics

36–53 53–80 80–100 100–155 155–313

Figure 1.1  Map of age-​standardized ischaemic heart disease mortality rate per 100,000 persons in 21 world regions, 2010; the Global Burden of Disease 2010 Study. Reproduced from Moran A et al. 'The Global Burden of Ischemic Heart Disease in 1990 and 2010'. Circulation (2014) 129:1483 with permission from the Wolters Kluwer Health, inc.

impacted middle-​income countries, where young adults were more likely to develop IHD (12, 15) (Fig 1.2).

Morbidity Coronary artery disease can present with a wide range of clinical syndromes, including stable angina, acute coronary syndrome, heart failure, arrhythmia, and death. Estimating the incidence and prevalence of coronary disease-​related morbidity is therefore FEMALE

28.8% 14,197,224

36.0% 17,750,676

MALE

29.4% 14,490,293

29.4% 23,004,851

5.9% 2,916,834

38.0% 29,765,638

High

Low

Lower-Middle

Upper-Middle

27.8% 21,719,473

4.8% 3,770,022

Figure 1.2  World Bank income category composition of absolute numbers of ischaemic heart disease (IHD) disability-​adjusted life years (DALYs) in males and females in 2010; the Global Burden of Diseases, Injuries and Risk Factors 2010 Study. Reproduced from Moran et al ‘Variations in ischemic heart disease burden by age, country, and income: the Global Burden of Diseases, Injuries, and Risk Factors 2010 study’. Global Heart 9:1 (2014) 91–​99 with permission from Elsevier.

challenging and is confounded by changing definitions and diagnostic criteria over time (16, 17). Acute coronary syndromes, including unstable angina and myocardial infarction (with and without ST-​segment elevation on the electrocardiogram), present a major health burden on industrialized societies. As with CHD mortality there are large regional, socioeconomic, and ethnic variations in the incidence and prevalence of myocardial infarction. The reported incidence and prevalence of myocardial infarction is higher in men than in women and increases with age. The Health Survey for England 2006 reported that 4.1% of all men and 1.7% of all women in the UK have had a myocardial infarct (5). In 2011 the prevalence of myocardial infarction in the UK was estimated to be 1.7% for men of all ages and 1% for women of all ages (3). In the USA in 2009–​12 the prevalence of myocardial infarction in adults aged 20 or over was estimated at 2.8%, with a prevalence of any coronary disease of 7.8% (18). In Scotland in 2000 there were over 9000 admissions to hospital with suspected acute coronary syndrome per million population, which accounted for 19% of all emergency hospitalizations and 12% of medical bed days (19). A  decrease in the ratio of ST-​elevation to non-​ST-​elevation myocardial infarction has been reported, but whether this is due to a real change in disease prevalence, an effect of treatment, or a change in case recognition is unknown (20). The incidence and prevalence of stable coronary disease is difficult to estimate. The incidence of angina in the UK is approximately 96,000 new cases a year, with a higher rate amongst men than women (5). In 2011 the prevalence of angina was estimated to be 3.9% amongst men and 2.5% amongst women (3).

Risk factors The INTERHEART study investigated various risk factors for myocardial infarction in 15,152 cases in 52 countries, who were matched

5

Chapter 1 

epidemiology and pathophysiology of coronary artery disease

to 14,820 controls with no history of heart disease. The mean age of first presentation with myocardial infarction was 8 years younger in men than women and 10 years younger in Africa, the Middle East, and South Asia than the rest of the world. Nine easily measured and potentially modifiable risk factors for myocardial infarction were identified, including smoking, hypertension, diabetes, waist to hip ratio, low daily fruit and vegetable consumption, physical inactivity, overconsumption of alcohol, abnormal blood lipid levels, and psychosocial factors. The effect of these risk factors was consistent in both genders and across different ethnic groups and geographic regions. Collectively, the nine risk factors accounted for 90% of the population-​attributable risk for myocardial infarction in men and 94% in women (21). These risk factors have been incorporated into a risk score, which has been validated in a large cohort from 21 countries (22). The risk factor burden is lower in low-​income countries than in middle-​or high-​income countries, but paradoxically the rates of major adverse cardiovascular events are higher in low-​income countries than in middle-​or high-​income countries. It has been suggested that the high-​risk factor burden in high-​income countries is mitigated by preventive medications and revascularization procedures, which are substantially more common in high-​income than in middle-​or low-​income countries (23). Tobacco use, perhaps the most important modifiable risk factor, is associated with a nearly threefold increase in the odds of myocardial infarction (odds ratio [OR] for current smokers 2.95, 95% confidence interval [CI] 2.77–​3.14 versus never smokers). This increase in risk of myocardial infarction falls after quitting smoking Women

(OR at 3 years 1.87, CI 1.55–​2.24) but remains elevated even after 20 or more years of abstinence (OR 1.22, CI 1.09–​1.37). These data suggest that the greatest reduction in global CHD risk could be achieved by preventing smoking and by smoking cessation programmes (24). A meta-​ analysis of data from 61 prospective observational studies involving almost 900,000 adults, mostly from Western Europe or North America, confirmed a strong positive relationship between total serum cholesterol and coronary mortality, irrespective of age and the level of blood pressure. Of various simple indices involving measurement of low-​density lipoprotein (LDL) and high-​density lipoprotein (HDL) cholesterol levels, the ratio total/​ HDL cholesterol was the strongest predictor of coronary mortality (25). Randomized trials of just a few years treatment with 3-​hydroxy-​3-​methylglutaryl-​coenzyme A  (HMG CoA) reductase inhibitors (statins) have shown that lowering LDL cholesterol by about 1.5  mmol/​L reduces the incidence of coronary events by about a third (26). Global reductions in other modifiable risk factors also have the potential to prevent cardiovascular events, but lowering rates of hypertension, obesity, and diabetes will be challenging. Epidemiological evidence suggests that throughout middle and old age usual blood pressure is strongly and directly related to vascular (and total) mortality without any evidence of a threshold down to at least 115/​75 mmHg (27). In the USA, however, from 2001 to 2003 state-​level age-​standardized prevalence of uncontrolled hypertension was estimated to range from 15% to 21% amongst men and from 21% to 26% amongst women (28) (Fig 1.3). Similarly, there Men

1988–1992

39.5 41.3 43.1 44.9 46.7

to to to to to

41.3 43.2 44.9 46.7 48.5

48.5 50.2 52.0 53.8 55.6

to to to to to

50.2 52.0 53.8 55.6 57.4

32.5 33.7 34.9 36.0 37.2

to to to to to

33.7 34.9 36.0 37.2 38.4

38.4 39.6 40.8 41.9 43.1

to to to to to

39.6 40.8 41.9 43.1 44.3

2001–2003

Figure 1.3  Age-​standardized prevalence (in percentage) of uncontrolled hypertension in the USA from 1988 to 1992 and from 2001 to 2003 (men and women ≥60 years of age). Hypertension control decreased in women between the two study periods. Reproduced with permission from Ezzati M, Oza S, Danaei G, Murray CJ. Trends and cardiovascular mortality effects of state-​level blood pressure and uncontrolled hypertension in the United States. Circulation 2008 Feb 19; 117(7):905–​14.

5

6

6

Section 1  

background and basics

is robust evidence that an increase in body mass index of 5 kg/​m2 is associated with about a 40% increase in vascular mortality (29), but from 1999 to 2006 the high prevalence of childhood obesity in the USA remained unchanged (30). Nevertheless, relatively modest downward shifts in the population distribution of modifiable cardiovascular risk factors may have substantial effects on disease prevalence, particularly when compared with treatment strategies directed at high-​risk individuals (31).

Pathophysiology Atherothrombosis Atherothrombosis, defined as atherosclerosis with superimposed thrombosis, is the principal pathological process underlying the majority of clinical cardiovascular events. Atherosclerosis is a systemic process that involves large and medium-​sized elastic and muscular arteries and typically affects the aorta and coronary, carotid, and peripheral vessels. The epicardial coronary arteries are particularly susceptible, but other arteries, including the intramyocardial arteries, are rarely affected. Atherosclerosis starts in childhood, progresses silently through early adult life, and often manifests in later decades with ischaemia or infarction of the heart, brain, or extremities. The disease is characterized by the development of focal atherosclerotic plaques within the intimal layer of the arterial wall that consist of cells, connective tissue, lipids, and debris. The cellular constituents include endothelial and smooth muscle cells from the vessel wall, and inflammatory and immune cells derived from the circulating blood. As the disease progresses individual plaque morphology may change abruptly because of plaque rupture and superimposed thrombosis. In addition, secondary changes may develop in the media and adventitia. As a consequence there may be marked heterogeneity in plaque morphology in different vascular territories, even in the same individual. The complex molecular and cellular mechanisms underlying the atherosclerotic disease process are incompletely understood, but it is now recognized that atherosclerosis is an active process involving interplay of cardiovascular risk factors, vascular biology, and chronic inflammation.

Endothelial activation The vascular endothelium, the innermost cellular layer of blood vessels, has a key role in vascular homeostasis and is critically involved in the development of atherosclerotic disease. In health the endothelium produces a wide range of locally active substances that regulate contractile, secretory, and mitogenic functions of the vessel wall, and influence blood coagulation.

Endothelial physiology The importance of the endothelium was first demonstrated in studies of vascular tone (32), but it is now recognized that the endothelium releases a range of autocrine and paracrine mediators that control vascular physiology and response to injury. Nitric oxide (NO), the principal endothelium-​derived relaxing factor, plays a key role in the maintenance of vascular tone and endothelial reactivity. NO is synthesized from the amino acid L-​arginine by the action of endothelial nitric oxide synthetase (eNOS). This enzyme requires a critical cofactor, tetrahydrobiopterin, to facilitate endothelial NO production. Following release from

endothelial cells, NO diffuses into medial smooth muscle cells and activates guanylate cyclase, which results in cyclic guanosine monophosphate (cGMP)-​mediated vasodilatation. In addition, NO maintains the endothelium and medial smooth muscle cells in a non-​proliferative state and, when released into the blood, NO inhibits platelets and leukocytes. An NO-​independent pathway also contributes to vasodilator tone but has not yet been fully elucidated (33–​35). The actions of NO are opposed by endothelium-​derived vasoconstrictor factors, such as endothelin and vasoactive prostanoids, and by angiotensin II, which is converted at the endothelial surface from angiotensin I. These mediators cause vasoconstriction, activate endothelial cells, platelets, and leukocytes, and facilitate thrombosis, directly countering the inhibitory effects of NO (33–​35).

Endothelial activation and dysfunction Exposure to cardiovascular risk factors (including tobacco use, hypertension, hyperlipidaemia, and diabetes) activates mechanisms within endothelial cells that result in expression of chemokines, cytokines, and adhesion molecules programmed to interact with leukocytes and platelets. At a molecular level risk factor exposure appears to induce a switch from NO-​mediated inhibition of endothelial and other cellular processes towards endothelial activation via redox signalling. As part of endothelial activation eNOS, which normally maintains the endothelium in a quiescent state via production of NO, switches to generate reactive oxygen species (ROS). This process is termed eNOS uncoupling, and results in superoxide production if there is tetrahydrobiopterin deficiency, and hydrogen peroxide production if levels of L-​arginine are inadequate. The resulting oxidative stress within the endothelium leads to increased production of endothelin and other mediators, which promote endothelial activation (34, 35). Collectively, these processes result in endothelial dysfunction, a systemic disorder affecting all arteries that predisposes to vasoconstriction, increased endothelial cell permeability, expression of adhesion molecules, increased chemokine secretion, leukocyte adherence and migration, vascular smooth muscle cell proliferation, and platelet activation and thrombosis (35, 36) (Fig 1.4). Clinical indicators of endothelial activation, such as endothelial vasomotor dysfunction, can predict cardiovascular events in patients with and without overt coronary artery disease (37) but correction of cardiovascular risk factors has been shown to improve endothelial function. For example, treatment of hypercholesterolaemia with statins has been shown to improve or normalize endothelial function in patients with mild coronary artery disease (38). Angiotensin-​converting enzyme inhibitors (ACE I) also improve endothelial function through a range of mechanisms (antioxidant effects, favourable effect on fibrinolysis, reduction in angiotensin II, increase in bradykinin), although a direct relationship between these effects and the risk of adverse cardiovascular events has not yet been clearly established (39).

Early stages of atherosclerosis The mechanisms that underlie the initial stages of atherosclerosis have not been fully elucidated but endothelial activation appears to be integral to the process. Endothelial activation precedes the onset of the disease, facilitates inflammatory processes that lead to atherosclerosis, and promotes mechanisms of disease progression.

7

Chapter 1 

A

epidemiology and pathophysiology of coronary artery disease

Monocytes

BLOOD

Lipoprotein Platelet activation

Platelets _

LDL

_

NO

+ Chemokines

CAMs

+

Endothelial activation

EC MCSF Ox-LDL

INTIMA

NO

ROS (O2–, OH–)

Foam cells

Endothelin

Elastic lamina SMC relaxation MEDIA

Macrophage differentiation & phagocytosis

SMC contraction

SMC

Figure 1.4  Simplified schematic of atherogenesis. Nitric oxide (NO) secreted by endothelial cells (EC) causes relaxation of smooth muscle cells (SMC) and vasodilatation. NO also inhibits (–​) platelets and leukocytes. Low-​density lipoprotein (LDL) enters the subendothelial space and is modified, generating oxidized LDL (Ox-​LDL). Endothelial activation and dysfunction causes generation of reactive oxygen species (ROS) and endothelin, expression of cell adhesion molecules (CAMs) on the endothelial cells, and activation of platelets and monocytes (+). Monocytes adhere to the endothelium and, under the influence of chemokines, migrate into the subendothelial space. Macrophage colony stimulating factor (MCSF) induces monocyte differentiation into macrophages. Activated macrophages phagocytose lipid and develop into foam cells.

Lipid retention and modification In the earliest stage of atherosclerosis LDL particles probably enter the subendothelial space from the bloodstream. Apolipoprotein in the LDL particles is thought to bind to extracellular proteoglycans (especially biglycan) and other macromolecules, ensuring retention of lipid within the extracellular matrix (40, 41). LDL particles may be modified through oxidation and glycation. The precise pathways of this chemical transformation are uncertain but evidence implicates myeloperoxidase and 12/​15-​lipoxygenase, peroxidase enzymes found predominantly in neutrophils, monocytes, and some macrophages (42, 43).

Inflammation Modified and oxidized LDLs contribute to endothelial activation and initiate an inflammatory response in the vessel wall. Activated endothelium expresses several types of cell adhesion molecules (CAMs), which facilitate adhesion of leukocytes rolling along the endoluminal surface of the vessel wall to the endothelium. Chemokines produced in the endothelial cells then stimulate migration of the adherent monocytes and T cell lymphocytes into the subendothelial space (44–​46). Macrophage colony stimulating factor, a cytokine produced in the activated endothelial cells, stimulates monocytes within the intima to differentiate into macrophages. Recruited macrophages express several different polarization phenotypes and have numerous effects on lesion development. The commonest phenotype is the M1 macrophage, which triggers a predominantly proinflammatory response (47). This M1 transformation is

associated with upregulation of scavenger receptors and Toll-​like receptors on the macrophage cell surface that bind modified LDL and oxidized phospholipid. Activation of macrophage Toll-​like receptors also induces intracellular signalling and cell activation, with cytoskeletal rearrangements, stimulation of inflammatory cytokine secretion, and production of proteases and cytotoxic oxygen radicals. These processes facilitate endocytosis and destruction of the oxidized LDL particles, but if the lipid cannot be fully metabolized it accumulates as cytosolic droplets and the macrophage transforms into a foam cell (44, 48). Other macrophages assume an M2 phenotype with predominantly anti-​inflammatory effects. This balance between proinflammatory and anti-​inflammatory phenotypes is incompletely understood but will have a significant influence on disease progression (49). Lymphocytes within the intima also produce inflammatory cytokines, chemokines, proteases, and cytotoxic oxygen and nitrogen radical molecules. Cytokines may induce expression of CD40, a transmembrane protein receptor present on inflammatory cells within the plaque. Activation of CD40 by CD40 ligand, derived from platelets and other cells, signals upregulation of proinflammatory and atherogenic genes (50). This process is known to involve the intracellular nuclear factor kappa B transcription pathway, which controls the transcription of genes for many cytokines, chemokines, adhesion molecules, and regulators of apoptosis (51). These processes augment and perpetuate the inflammatory atherosclerotic process and recruit additional macrophages and medial smooth muscle cells. If the inflammatory response does not remove or neutralize the initiating stimulus it can continue unabated.

7

8

8

Section 1  

background and basics

The accumulation of lipid-​laden monocytes, foam cells, and T cell lymphocytes within the intima leads to the formation of fatty streaks and early atherosclerotic lesions (Fig 1.5). Fatty streaks are prevalent in young people and are generally considered to be an antecedent of atheroma, but they may also disappear over time (52). Evidence of early atherosclerosis has been demonstrated in post-​mortem studies of young soldiers killed during the Vietnam (53) and Korean (54) wars and in intracoronary ultrasound studies of transplanted hearts retrieved from teenage and young adult donors (55).

Disease progression Plaque growth As the atherosclerotic process progresses the plaque increases in size due to accumulation of inflammatory and smooth muscle cells, production of extracellular matrix, and continuing deposition of lipid in the arterial wall. Vascular smooth muscle cells, stimulated by mitogens and cytokines, differentiate into migratory and secretory cells and migrate into the intima (56) (Fig 1.6) Smooth muscle cells produce collagen and other matrix proteins, including glycosaminoglycans, proteoglycans, elastin, fibronectin, laminin, vitronectin, and thrombospondin (57).

Arterial remodelling During growth of atherosclerotic plaque the entire vessel can vary in size, a process known as remodelling. Enlargement of the vessel may accommodate the plaque volume without compromising the arterial lumen until the plaque enlarges to over 40% of the vessel cross-sectional area, but thereafter further growth in the plaque causes luminal narrowing (57, 58). Alternatively, the vessel may constrict and further narrow the arterial lumen. Progressive luminal narrowing can obstruct coronary blood flow and lead to stable angina pectoris. The mechanisms regulating remodelling

Intimal thickening

have not been elucidated but may contribute to heterogeneity in the progression and clinical manifestations of arterial disease (59).

Plaque neovascularization As atheromatous disease advances, new microvessels may develop from the adventitial vasa vasorum, possibly in response to hypoxia and activation of Toll-​like receptors within the expanding atherosclerotic plaque. This process appears to be regulated by vascular endothelial growth factor (VEGF) A, which, together with angiotensin II, can also induce microvascular permeability. These processes may facilitate extravasation of red blood cells and intraplaque haemorrhage. Release of haemoglobin into the extracellular matrix exacerbates oxidative stress, amplifying macrophage activation and proinflammatory signals, and accelerating the atherosclerotic process (60).

Apoptosis Apoptosis of the cellular components of the plaque may be mediated by cytokines, including interleukin-​1, tumour necrosis factor-​ alpha, and interferon-​gamma (61). Apoptosis has been observed at all stages of atherosclerosis but the consequences for lesion progression may depend on how efficiently the apoptotic cell is cleared by other macrophages. This phagocytic clearance (efferocytosis) appears to be efficient in early lesions, reducing lesion cellularity and atheroma progression. In more advanced lesions efferocytosis may be defective, leading to secondary necrosis of the apoptotic cell, further release of inflammatory mediators, and amplification of the inflammatory process (62). Cumulatively these events may lead to the development of a highly thrombogenic necrotic core within the expanding plaque, which contains cell remnants expressing active tissue factor (63). As the necrotic lipid-​rich core expands, a fibrous cap forms over the luminal surface, creating a barrier between the thrombogenic material within the core and the circulating blood (Fig 1.6).

Pathologic intimal thickening

Intimal xanthoma

Fibrous cap atheroma

Thin-cap Fibroatheroma NC

NC

EL

FC

Erosion

Rupture

Calcified nodule

NC

Healed Rupture

Nodule Th

Th

NC

Th

Th FC

Ca 2+

Figure 1.5  Histopathology of plaque progression. Descriptions begin at top, from left to right. Intimal thickening is normal in all age groups and is characterized by smooth muscle cell accumulation within the intima. Intimal xanthoma corresponds to the fatty streak and denotes the accumulation of macrophages and lymphocytes within the intimal thickening lesion. Pathological intimal thickening denotes the accumulation of extracellular lipid. Fibrous cap atheroma indicates the presence of a necrotic core under a fibrous cap, which may become thinned (thin-​cap atheroma). This lesion may rupture, with exposure of the necrotic core to the lumen. The thrombus of a plaque erosion may overlie pathological intimal thickening (left) or fibrous cap atheroma (right). Calcified nodule is a rare cause of coronary thrombosis. Acute rupture may progress to healing (healed plaque rupture) without luminal occlusion. EL, Extracellular lipid; FC, fibrous cap; NC, necrotic core; Th, thrombus. Reproduced from Frostegard J, Ulfgren AK, Nyberg P, et al. Cytokine expression in advanced human atherosclerotic plaques: dominance of pro-​inflammatory (Th1) and macrophage-​stimulating cytokines. Atherosclerosis 1999; 145(1):33–​43 with permission from Elsevier.

9

Chapter 1 

epidemiology and pathophysiology of coronary artery disease

BLOOD Circulating endothelial cells

Monocyte recruitment

Vulnerable plaque with thin fibrous cap Apoptosis

Chemokines CAMs

Proteases

Matrix protein Cytokines Apoptosis

SMC

INTIMA Ox-LDL accumulates

MMPs

Mitogens

Matrix degradation

SMC

Necrotic core

SMC migration & differentiation

MEDIA

SMC

Figure 1.6  As oxidized lipid accumulates, monocytes are recruited to the developing plaque. Cytokines and mitogens stimulate recruitment and proliferation of smooth muscle cells (SMCs). SMCs produce extracellular matrix, which increases plaque volume. Apoptosis of endothelial cells and impaired endothelial regeneration may lead to plaque erosion. Apoptosis of cells within the plaque leads to the development of a lipid-​rich necrotic core. The overlying fibrous cap may be degraded by matrix metalloproteinases (MMPs) and other proteases, increasing the risk of plaque rupture. Other abbreviations as in Fig 1.4.

Endothelial cells can also progress to senescence and may detach into the circulation. Whole endothelial cells and microparticles derived from activated or apoptotic endothelial cells can be detected in the circulating blood as markers of endothelial injury and are thought to influence blood thrombogenicity (64). Restoration of endothelial integrity involves replication of adjacent mature endothelial cells or recruitment of circulating endothelial progenitor cells. Mobilization of endothelial progenitor cells is influenced by NO and may therefore be impaired in individuals with cardiovascular risk factors (35, 65). In animal models restoration of endothelial integrity after balloon injury is enhanced with exercise or statins, which both improve endothelial function (66, 67).

Influence of biomechanical forces Dysfunctional endothelium, fatty streaks, and atheroma all localize preferentially to arterial sites associated with disturbed flow patterns, suggesting an important role for local haemodynamic forces in the development of arterial disease. These sites include branch points on the opposite side of the flow divider, and post-​stenotic segments where disturbances in laminar flow result in recirculation eddies, flow separation, and oscillatory flows. Evidence suggests that exposure of the endothelium to such different biomechanical forces induces differential expression of specific genes in endothelial cells. Laminar shear stress from the viscous flow of blood against the endothelial cell surface induces eNOS activity, which supports vasoprotective functions in the endothelium. By contrast, reduced or oscillatory shear stress induces endothelial activation, expression of adhesion molecules, and endothelial cell apoptosis (68–​73).

Calcification of atheroma Microscopic areas of calcification may appear within the atherosclerotic plaque, which become denser as the disease advances. The extent of coronary calcification correlates closely with the severity of luminal narrowing caused by the plaque (74). The predominant chemical constituent of coronary calcification is identical to hydroxyapatite, the main inorganic constituent of bone (75). Osteopontin, a gylcosylated protein involved in the formation and calcification of bone, is synthesized by macrophages, smooth muscle, and endothelial cells. Endothelial progenitor cells in patients with coronary disease have also been shown to express osteocalcin, an osteoblastic marker (61, 76). The significance of calcification for plaque progression and cardiac events is uncertain, but extensive calcification may impact the outcome of percutaneous coronary intervention. Rarely, eruptive nodular calcification with underlying fibrocalcific plaque is implicated as a cause of coronary thrombosis (77) (Fig 1.5).

Plaque rupture, erosion, and thrombosis Encroachment of atherosclerotic plaque into the lumen of a coronary artery without thrombosis can cause stable angina pectoris; however, acute coronary syndromes are caused by luminal thrombosis or sudden plaque haemorrhage into an atherosclerotic plaque with or without concomitant vasospasm. In ST elevation myocardial infarction (STEMI) the thrombus is occlusive and sustained, whereas in unstable angina and non-​ST-​elevation myocardial infarction (NSTEMI) the thrombus is typically non-​occlusive and dynamic (78). Detailed histopathological examination of coronary arteries in sudden cardiac death victims has revealed two broad categories of plaque ‘events’ that lead to thrombosis, with

9

10

10

Section 1  

background and basics

approximately 75% of fatal coronary thrombi due to plaque rupture and the remaining 25% due to plaque erosion (77).

Plaque rupture Most atherosclerotic plaques develop slowly over many years, under the influence of local immune responses and continued exposure to cardiovascular risk factors. Integrity of the fibrous cap overlying the plaque core is maintained by balanced production and degradation of extracellular matrix proteins. If this balance is disturbed, overproduction of matrix may encroach on the arterial lumen, but increased matrix degradation may weaken the plaque cap, increasing the risk of plaque rupture. Matrix protein degradation is mediated by matrix metallo­ proteinases (MMPs) and other proteases released by inflammatory cells, including macrophages and migrated smooth muscle cells. MMPs are zinc2+-​ dependent endopeptidases and include collagenases, gelatinases, stromelysins, and metalloelastases. MMP activation is controlled at several levels, including induction of MMP gene transcription, post-​translational activation of MMP proforms, and interaction with specific tissue inhibitors (TIMPs). MMPs may facilitate smooth muscle cell migration through the internal elastic lamina into the intima, are implicated in vascular remodelling, and appear to have a central role in plaque rupture. Expression of MMP activity is influenced by several drugs, including the HMG CoA reductase inhibitors (statins) (57, 79). Active degradation and remodelling of the extracellular plaque matrix by macrophages, via release of MMPs and other proteases and by subsequent phagocytosis, inhibits the formation of a stable fibrous cap. Further breakdown of collagen and other proteins within the fibrous cap reduces the structural integrity of the plaque and predisposes to plaque rupture (79, 80). Interaction between CD40 and CD40 ligand may induce MMP production and may play a role in plaque instability (50). Plaques with a thin fibrous cap, large lipid core, and inflammatory cell infiltrate at the thinnest portion of the cap appear to be particularly vulnerable to rupture (77) (Fig 1.7). Inflammatory cells are abundant in the shoulder regions of ruptured plaque and many show signs of activation and inflammatory cytokine production (81, 82). Histopathological specimens show rupture typically at this thin shoulder region of the collagenous fibrous cap, with discontinuity of the cap at the site of contact between thrombus and the lipid core. In one study 95% of ruptured plaques had a cap thickness of 50% diameter stenosis in three vessels), but assessment by measurement of fractional flow reserve may demonstrate only one to be obstructive and requiring intervention (9), and there is evidence that revascularization guided by such information leads to improved outcomes (10). As clinicians understand the basis of the variables used for any particular risk model, so there is the possibility both of risk-​averse behaviour and of ‘gaming the model’. If a patient is recognized as being at higher risk of an adverse outcome than is defined in the known measured parameters (for example, coronary vessels look difficult to graft), there is an anxiety that the actual outcome will be worse than that predicted. The operator will perform less well than the model even if performance is in fact satisfactory. This is discussed in greater detail below in the section on ‘Problems with publication of outcome data’. Subjective features in a model that increase the predicted risk can be overemphasized, which will tend to make predicted outcomes worse, and observed outcomes relatively better than the model. The accuracy of the data entered in the dataset is critical to the validity and reliability of resulting models. Mortality is the easiest endpoint to measure and validate and as a result is most often included. Many large datasets include in-​ hospital major adverse cardiovascular and cerebrovascular events (MACCE), and thus several models can estimate risk of this combined outcome. There have been attempts to determine whether data routinely collected for administrative rather than clinical purposes could be used to generate reliable risk models. Hannan compared the administrative data on the Statewide Planning and Research Cooperative System with the clinical Cardiac Surgery reporting System to assess unit-​specific mortality following coronary artery bypass surgery in New York State and found the administrative database was considerably inferior, mainly due to three clinical risk factors (ejection fraction, reoperation, and left main stem disease) (11). Ugolini and Nobilio (12) compared administrative datasets with EuroSCORE and found that, by adding administrative variables considered a proxy for clinical complexity to the administrative dataset and linking data across multiple episodes of patient care, they were able to eliminate much of the difference between the clinical and administrative dataset predictions. Certainly there has always been an interest in using routinely collected administrative data to this end, and some have argued that, although the data quality in administrative datasets is often very unreliable, once they start to be used in this way and clinicians engage with the process of data collection and validation, the quality of the data will improve and they may become a valuable way to model outcomes. This may not be particularly relevant in cardiology, where a long tradition of gathering clinical audit data exists, but it may have a role in subspecialties where a focus on data collection has been less prominent, and also the assessment of the outcomes following a clinical presentation (e.g. all myocardial infarction) rather than a specific procedure (e.g. PCI).

Risk adjustment models: creation and assessment Following selection of a dataset, statistical manipulation will generate a model. Bayesian methods have been used, but most models are generated using logistic regression analysis. This technique predicts the binary outcome, such as death or survival, on the basis

risk assessment and analysis of outcomes

of variables contained within the dataset. These predictor variables can take any form. They can be continuous, discrete, or dichotomous, and no assumptions are made about their distribution. The variables also need not be independent, though the additional predictive value of a variable already correlated with one in the model will be less. The dataset is initially analysed using univariate methods to find individual variables that predict outcome. Univariate analysis refers to the simple description of a set of values to assess their central tendency and spread. For example, in a normally distributed group of continuous variables, this might be the mean and standard deviation (SD). Logistic regression is then used to create a model using the identified variables. The goal of logistic regression is to correctly predict the category of outcome for individual cases using the least number of variables. There are a variety of methods used to try to find the minimum number of predictors that give most discriminatory power for a model. Broadly, the variable with the biggest predictive power is added to the model, then multivariate analysis used to find the next most important. An alternative technique involves the creation of a model that includes all the variables, and then variables are sequentially removed to see if their absence reduces the model’s discriminatory power. The selection of predictor variables is more than a statistical decision. Variables that can be measured objectively are favoured over subjective ones, though relatively important subjective features may still need to be included. If the risk model will be used to try to predict outcomes, only factors that can be measured before the intervention should be included. Some promising data fields may have low levels of data completeness, and then the options lie between omitting them from the model or trying to impute missing values. Sensitivity analyses can be performed to try to gauge how deleterious are the missing data on the model’s performance. The model takes the form of a mathematical equation that, when provided with the specific characteristics of the patient about to have the PCI, will generate a number, which is the predicted chance of the adverse outcome in question occurring (for example, the chance of death at 30  days). Once a model has been created it is tested to see how well it performs in predicting outcomes. This is done by assessing two critical features: calibration and discrimination. Calibration assesses the agreement between observed and predicted outcomes across the entire spectrum of predicted risk. Discrimination measures how well the model correctly separates those that go on to have the outcome from those who do not.

Calibration Calibration is often assessed with the Hosmer-​Lemeshow technique (13). The dataset is usually divided into 10 groups, based on the predicted outcome, ranging from low risk (P = 0 to 0.1) to high risk (P = 0.9 to 1.0). It is important that the dataset used has appropriately large numbers of patients represented at each level of risk. The observed outcomes at each risk level are then compared with the model-​predicted outcomes, a good model being one that gives an accurate prediction of outcome at each level of risk. Alhough the chi-​squared statistic is used, it may be unreliable in large datasets (14) and so visual inspection of the calibration plots is also important. A P value of >0.1 usually indicates that the model provides a good fit for the data and that the differences are not statistically significant, but nevertheless does not exclude potentially clinically significant differences between observed and predicted outcomes.

27

28

background and basics Threshold (drag left or right)

60 50 40 30 20 10

0

0.1

0.2

0.3

0.4 0.5 0.6 0.7 Risk predicted by model Negatives

0.8

0.9

1

Positives

Figure 3.1  ‘Negatives’ are patient who do not experience an adverse outcome, and ‘positives’ are those that do.

Discrimination Discrimination assesses the positive and negative predictive accuracy of the model and is usually assessed using receiver operating characteristic (ROC) curves (15, 16). To help explain

At Threshold 0.8

60 50

True Positive Rate 121/548 = 0.22

40 30

False Positive Rate 0/540 = 0

20 10 0

Plot on ROC curve 0, 0.21 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

60

FPR

At Threshold 0.5

50

True Positive Rate 529/548 = 0.97

40 30 20

False Positive Rate 293/540 = 0.54

10

Plot on ROC curve 0.54, 0.97

0

ROC curve

TPR

0

the principle, first recall that the ‘sensitivity’ of a test is also known as the ‘true positive rate’ (TPR), which is the proportion of patients with adverse outcome that are correctly identified. ‘Specificity’ is the ‘true negative rate’ and measures the proportion of patients without an adverse outcome correctly identified. The ‘false positive rate’ (FPR) is 1—​specificity, and is the proportion of patients without an adverse outcome incorrectly identified. An ROC curve is a graph of TPR on the y axis against FPR on the x axis, generated for the entire range of cut-​off values that are used to classify a patient as negative (no adverse outcome) or positive (with an adverse event). A visual representation of these principles is provided in Fig 3.1. In this example a total of 1088 patients is shown. Every blue unit represents a patient without an adverse outcome (a total of 540), and every red unit a patient with an adverse event (total 548). The x axis shows the model’s predicted outcomes, and the y axis the count of patients according to observed outcome. So when the model predicted a chance of 0.3 that a patient would have an adverse event, the dataset shows 10 patients and all were correctly identified (all blue, none red). When the chance was predicted to be 0.5, 50 patients were event-​free (blue) and 10 experienced an adverse outcome (red). Now consider the threshold (Fig 3.1). Red patients to the right of the line are the correct predictions, so the TPR is calculated as the number of these patients divided by all red patients. Blue patients to the right of the line are the false positives, so the FPR is the number of blue patients to the right of the line divided by all of the blue patients. To create an ROC curve, the TPR and FPR is found for every predicted level. Fig 3.2 shows how these curves are plotted.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Figure 3.2  The creation of a receiver operator curve (ROC). FPR, False positive rates; TPR, true positive rates.

TPR

Section 1  

Number of patients

28

FPR

29

Chapter 3 

60

risk assessment and analysis of outcomes ROC curve

50

TPR

40 30

Area ≈ 0.5

20 10 0

FPR 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

60 50

TPR

40 30

Area ≈ 1

20 10 0

FPR 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Figure 3.3  How a receiver operator curve (ROC) demonstrates the discriminant power of a model. FPR, False positive rates; TPR, true positive rates.

The c statistic is simply the area under the ROC curve. If the model is poor, it will be unable to discriminate between patients who will or will not have an adverse outcome, the red and blue patients will be almost superimposed, and the ROC will approach a straight line (Fig 3.3). The area under this curve will approach 0.5, and the model is no better at predicting outcome than the toss of a coin. A good model will be able to separate the patients, and push the curve up, towards perfect discrimination, which gives an area of 1. Logistic regression models with a c statistic of about 0.8–​0.9 are usually regarded as having adequate clinical utility. Important limitations must be recognized with the use of these general statistical methods when they are applied to predict outcome in individual patients. A ‘good’ model with a c statistic of 0.8 will still miss 20% of patients with an adverse event. Furthermore, even if a model has a perfect fit, the prediction of a 10% mortality in 1000 patients does not highlight the 100 who will die. Box-​Jenkins, an industrial statistician, stated that ‘all models are wrong, but some are useful’. We do well to bear this in mind.

Models for adult intervention and revascularization Adult cardiac surgery In cardiothoracic surgery, EuroSCORE (17) became the dominant method for calculation of perioperative mortality in the 2000s. Developed from a large clean dataset of a heterogeneous group of patients, the authors created a simple additive score. At a time when computers were far less accessible, it could be readily calculated at the bedside. An alternative based on the Society of

Thoracic Surgeons (STS) database in the USA was unavailable for critical analysis owing to intellectual property issues, and so gained less widespread use. The EuroSCORE has been tested in widely differing populations both in Europe (18), Scandinavia (19), Japan (20), and in the USA (21), and has consistently worked well despite varying patient demographics and operative techniques. Patients in the USA were older, more often had isolated coronary bypass operations, diabetes (30% versus 17%) and prior cardiac surgery (11% versus 7%), but EuroSCORE still outperformed the STS model (21). Across Europe EuroSCORE was reliable even though there were particularly large differences in valve surgery (ranging from 18.6% in Finland to 51.5% in Spain) (18). Although the development of the additive version of EuroSCORE greatly widened its applicability, one of the problems of trying to generate a simple scoring system from the underlying logistic regression equation is that it forces an additive system on one that is essentially multiplicative. In general, as the number of risk factors rises, the logistic score will tend to exceed the additive. This has the effect of making the additive score underestimate risk for high-​risk patients, and overestimating risk for low-​risk patients. This effect on the additive EuroSCORE has been well documented (22). With increasing computing power available at the bedside the full logistic EuroSCORE became more widely used. Limitations of the model in certain patient subsets has been documented—​the score is poor at differentiating risk in high-​risk patients undergoing CABG (23) and aortic valve replacement (24). Unfortunately, it is this minority of high-​risk cases that will have the greatest impact on a surgeon’s overall measured performance.

29

30

30

Section 1  

background and basics

Perhaps of more fundamental concern, and a recurrent difficulty in this field, is that risk adjustment models become outdated. As the demographic features of patients who are treated shifts, and techniques and associated pharmacology evolve, so the factors that are associated with adverse outcome and the strength of those associations changes. The very process of risk score creation enforces scores to be out of date by the time they are used, because they are generated using data from historic cohorts. Even 10 years ago it was documented that surgeons consistently outperformed both the additive and logistic EuroSCORE models (25). In 2012 an updated model called EuroSCORE II was published (26), and details, including a calculator, are available on the EuroSCORE website (http://​www. euroscore.org/​). EuroSCORE II used a similar methodology to the original but was based on a more recent cohort of operations (performed in 2010), with some changes to variables used in the model and recalibration. For example, left ventricular function is now described with the additional category of ‘very poor’ (left ventricular ejection fraction [LVEF] 20% or less), raised pulmonary artery systolic pressure—​previously dichotomized at greater than 60 mmHg—​ is now split into two groups (30–​55 and 56 and above), and creatinine clearance is used as a continuous variable. EuroSCORE II has been assessed in a number of surgical series. In the UK it performed well but was poorly calibrated for isolated CABG, and also in the lowest and highest risk patients (27). In Italy it was found to have good discrimination, and was well calibrated for the majority of patients; however, for the minority at very high risk (>30% predicted mortality) the model overpredicted adverse outcomes (28). Others have also found poor performance for higher risk cohorts (29). As the age of the population being treated increases, so frailty becomes important in defining risk. Frailty is theoretically defined as a clinically recognizable state of increased vulnerability resulting from ageing-​ associated decline in reserve and function across multiple physiological systems such that the ability to cope with everyday or acute stressors is reduced (30). The addition of frailty to the STS score has recently been shown to improve the model’s discrimination (31).

Transcatheter aortic valve implantation Patients being treated by TAVI represent a particularly elderly cohort, with a mean age of 81.3 years in the UK (32). There have been attempts to derive models for patients being treated but none currently perform well enough to inform doctors or patients adequately of the relative risks of surgical or the newer alternatives (33, 34). Frailty measured using the Katz Index (35) has been shown to be associated with adverse early and late outcomes after TAVI (36), and the inclusion of the Clinical Frailty Scale (37) has been shown to improve model discrimination (38). Nevertheless, frailty measured by the Geriatric Status Scale (39) was not associated with 30-​day mortality on multivariate analysis in the Italian OBSERVANT registry (33).

Percutaneous coronary intervention By comparison, numerous risk assessment systems have been developed for PCI and, owing to the rapid evolution of PCI technology and adjunctive pharmacological therapy, the more recent models have greater applicability to contemporary PCI. While most models were created to apply to the full spectrum of patients treated (40–​47), some have focused on a single clinical syndrome, such as ­primary PCI for ST segment elevation myocardial infarction

(STEMI) (48, 49). Kunadian et al. (50) assessed the Mayo Clinical Risk Score (47) and the North West Quality Improvement Program score (42) in a cohort of over 5000 patients treated between 2002 and 2006 in north-​west England. Both models had excellent discrimination and calibration, with area under ROC curve (c index) of ≥0.87 and 0.86, respectively. More recent models have successfully focused on accurate risk prediction in the highest risk cohorts, including those in cardiogenic shock and those who have sustained cardiac arrest prior to PCI (51). Because there is considerable overlap of features that predict risk both from CABG and PCI, EuroSCORE can risk-​stratify for percutaneous revascularization (52), and PCI risk scores can predict CABG outcome (53). Certain issues, however, are unique to the treatment modality. For PCI the specific characteristics of the lesions treated can weigh heavily in the likely chance of procedural success or adverse outcome. Technological improvements have removed some of the impact of lesion anatomy on early outcomes, but these have been counteracted by the increasing complexity of lesions tackled, which may fare less well in the longer term. Lesion-​ specific issues have been addressed in some models (54, 55), but the description of lesion characteristics has been relatively crude. The SYNTAX score (56) characterizes the coronary vasculature in more detail with respect to number of lesions and their functional impact, location, and complexity. The score is based on modifications and additions to a number of existing systems. The arterial tree is divided into 16 segments, using the classification from the Arterial Revascularisation Therapies Study (ARTS) I  and II (57) trials (Fig 3.4), and the contribution of each segment to total left ventricular bloodflow is used as a multiplication factor depending on left or right dominance (Table 3.1). Only vessels larger than 1.5  mm in diameter are included and a lesion is defined as a 50% or more reduction in luminal diameter. The only distinction in lesion severity is between occlusive and non-​occlusive stenosis, with a multiplication factor of 2 for non-​ occlusive lesions and 5 for occlusive lesions, reflecting the difficulty of the percutaneous treatment for the latter. All other adverse lesion characteristics have an additive value (Table 3.2). Tandem lesions are considered a single lesion if they are separated by less than three lesion reference diameters. To classify bifurcation lesions, a synthesis of the Duke (59) and Institut Cardiovasculaire Paris Sud (ICPS) (60) classification systems was employed, and is shown in Fig 3.5. Thus the SYNTAX classification has seven types, whereas the previous two systems had six types each. SYNTAX type G was missing from the Duke system, and type D was missing from the ICPS system. In addition, side branch angulation is included in the score on the basis that decreasing angulation (between side branch and distal limb of bifurcation) makes stent coverage of the ostium harder to achieve. Bifurcations are only considered for segment junctions: 5/​6/​11, 6/​7/​9, 7/​8/​10, 11/​13/​12a, 13/​14/​14a, and 3/​4/​16, and 13/​14/​15 in case of left dominance. Trifurcations are weighted as in Table 3.2, and only scored for segment junctions 3/​4/​16/​16a, 5/​6/​11/​12, 11/​12a/​12b/​13, 6/​7/​9/​ 9a and 7/​8/​10/​10a. Aorto-​ostial lesions occur in the left main stem (LMS) (segment 5) and right coronary artery (RCA) (segment 1), unless the left coronary has separate origins, in which case segments 6 and 11 may also score as aorto-​ostial. Diffuse disease is a parameter introduced to reflect increase difficulty in the creation of a surgical anastomosis, and goes some way to addressing the ‘graftability’ in the section on Adjustments for case mix discussed

31

Chapter 3 

risk assessment and analysis of outcomes

Left dominance 5

1

11

12

9

6

9a

12a

13 2

14

7

10

12b

10a

14a 3

14b

8

15

Right dominance 1

5 6

11

12a

13 2 16 3

16c 16b 4 16a

9 9a

12

14

7

10

12b

10a

14a 14b

8

Figure 3.4  Definition of the segments of the coronary artery tree. LAD, Left anterior descending (coronary artery); LCA, left coronary artery; LCX, left circumflex coronary artery; RAO, right anterior oblique; RCA, right coronary artery. Reproduced from Sianos G, Morel M-A, Kappetein AP, et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroInterv. 2005;1:219–​27, with permission from Europa Edition.

above. It is defined as present when at least 75% of the length of the segment distal to the lesion has a vessel diameter of 1.5  mm diameter (bifurcations or trifurcations). Finally, the last question is the only one that does not relate to each lesion, since it relates to anatomy beyond the stenosis. It is therefore scored once only for each major arterial territory (LMS, LAD, circumflex [Cx], and RCA). The final total SYNTAX score is the summation of all the scores for each lesion. Two examples of the calculation of this score are reproduced from Sianos el al. (56) in Fig 3.6. The hypothesis that higher SYNTAX scores, indicative of more complex disease, represent a bigger treatment challenge and are associated with poorer outcomes, has been confirmed. Initial analysis of performance in predicting short and longer term risk in patients with multivessel disease was encouraging (44). Patients with multivessel disease treated by PCI in the ARTS II study who were in the highest tertile of SYNTAX score (>26) had a significantly

31

32

32

Section 1  

background and basics

Table 3.1  Segment weighing factors. Based on the ‘Leaman’ score (58). Reprinted from Sianos G, Morel M-​A, Kappetein AP, et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroInterv. 2005;2:219–​27, with permission from Europa Edition.

Table 3.2  Segment weighing factors (SYNTAX scoring system). Reprinted from Sianos G, Morel M-​A , Kappetein AP et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroInterv. 2005;2:219–​27, with permission from Europa Edition.

Segment Segment name number

Right dominance

Left dominance

Lesions adverse characteristic scoring

1

RCA proximal

1

0

2

RCA mid

1

0

3

RCA distal

1

0

4

Posterior descending artery

1

n/​a

16

Posterolateral branch from RCA

0.5

n/​a

16a

Posterolateral branch from RCA

0.5

n/​a

16b

Posterolateral branch from RCA

0.5

n/​a

16c

Posterolateral branch from RCA

0.5

n/​a

5

Left main

5

6

6

LAD proximal

3.5

3.5

7

LAD mid

2.5

2.5

Trifurcations

8

LAD apical

1

1

1 diseased segment

+3

9

First diagonal

1

1

2 diseased segments

+4

9a

First diagonal a

1

1

3 diseased segments

+5

10

Second diagonal

0,5

0.5

4 diseased segments

+6

10a

Second diagonal

0.5

0.5

Bifurcations

11

Proximal circumflex artery

1.5

2.5

Type A, B, C

+1

12

Intermediate/​anterolateral artery

1

1

Type D, E, F, G

+2

12a

Obtuse marginal a

1

1

Angulation 20 mm

+1

14a

Left posterolateral a

0.5

1

Heavy calcification

+2

14b

Left posterolateral b

0.5

1

Thrombus

+1

15

Posterior descending

n/​a

1

‘Diffuse disease’/​small vessels

+1/​per segment number

Diameter reduction* Total occlusion

×5

Significant lesion (50–​99%)

×2

Total occlusion (TO) Age >3 months or unknown

+1

Blunt stump

+1

Bridging

+1

First segment visible beyond TO

+1/​per non-​visible segment

Side branch (SB)

Yes, SB 33). In all, 1800 patients were randomly assigned to CABG (n = 897) or PCI (n = 903) using a first-​generation paclitaxel-​eluting stent. Additional information was derived from registries nested within the trial. After 5  years follow-​up MACCE was higher following PCI than CABG (37.3% versus 26.9%, P < 0.0001) (61). This was mainly driven by repeat revascularization (25.9% versus 13.7%, P < 0.0001). All-​cause death was not statistically different (PCI 13.9% versus CABG 11.4%, P = 0.1) and neither was stroke (PCI 2.4% versus CABG 3.7%, P = 0.12). However, there were important differences in outcomes according to

* In the SYNTAX algorithm there is no question for % luminal diameter reduction. The lesions are considered as significant (50–​99% luminal diameter reduction) or occlusive. ** If all the side branches are 1.5 mm in diameter, no points are added since the lesion is considered as a bifurcation and will be scored as such.

SYNTAX score. Only in the lowest tertial was overall MACCE equivalent with both treatment modalities. Many have argued that repeat revascularization by PCI is a relatively benign event and should not be afforded equal weight as the other major adverse outcomes. For mortality, there was equivalence between PCI and CABG in lowest and intermediate tertiles, but in the highest tertile, with SYNTAX scores ≥33, PCI was associated with increased mortality. This trial started a new era in the objective assessment of coronary artery disease complexity, and SYNTAX score is now incorporated into European (62) and US revascularization guidelines (63). The combination of anatomical and clinical features might be expected

3

Chapter 3 

A

B

risk assessment and analysis of outcomes C Parent vessel only

Prebranch

Postbranch

D

F

E Bifurcation

Ostial

Prebranch & Ostial

G

Postbranch & Ostial Type A: Type B: Type C: Type D: Type E: Type F: Type G:

Pre-branch stenosis not involving the ostium of the side branch. Post side branch stenosis of the main vessel not involving the origin of the side branch. Stenosis encompassing the side branch but not involving its ostium. Stenosis involving the main vessel and ostium of the side branch. Stenosis involving only the ostium of the side branch. Stenosis directly involving the main vessel (pre-side branch) and the ostium of the side branch. Stenosis directly involving the main vessel (post-side branch) and the ostium of the side branch.

Figure 3.5  Institut Cardiovasculaire Paris Sud classification of bifurcations. Reproduced from Sianos G, Morel M-A, Kappetein AP, et al. The SYNTAX score: an angiographic tool grading the complexity of coronary artery disease. EuroInterv 2005; 1:219–27, with permission from Europa Edition.

to be an even more powerful way to allocate optimal treatment strategy and predict outcomes. Within any one tertile it is possible that clinical features will identify a subset at higher or lower risk. The SYNTAX score II was developed to address this issue (64). Using the patients in the SYNTAX trial with follow-​up to 4 years, two anatomical features (SYNTAX score and unprotected left main stem disease) were combined with six clinical variables (age, creatinine clearance, LVEF, sex, chronic obstructive airways disease, and the presence of peripheral vascular disease). Of note, diabetic status was not an independent predictor of outcomes in SYNTAX and so was not included. It can be seen that the SYNTAX score makes no contribution to the predicted outcomes from surgery. The resulting model was externally validated in the Drug Eluting stent for LefT main coronary Artery disease (DELTA) registry (65). A nomogram (Fig 3.7) was created that allows the bedside prediction of 4-​year mortality for an individual patient being treated either by PCI or CABG. The SYNTAX score II was indeed able to identify patients within any one SYNTAX score tertile at differing risk. Patients who were younger, female, and with lower ejection fraction required lower SYNTAX scores to achieve similar 4-​year outcomes between PCI and CABG, and those who were older, with pulmonary disease and left main stem lesions, could have higher SYNTAX scores, thus more complex anatomical disease for similar outcomes.

Assessing and displaying outcome data With the models described above in the section on ‘Models for adult intervention and revascularization’, we are now in a position

to use them to calculate predicted outcome following an intervention, compare this with observed outcome, and so try to judge if the quality of healthcare being delivered is satisfactory. The aim is to identify potential problems and focus on areas for improvement.

Periodic assessment Periodic, also known as cross-​sectional or static, assessment compares what would be expected from the model with what was observed at a fixed time period. Results are, by definition, retrospective although there are advantages as there is time for the data to be cleaned and validated. Data may be presented as crude survival or mortality plots, but by using risk models they can also be presented as a comparison between observed and expected numbers of deaths, sometimes expressed as a ratio with a confidence interval to reflect random variation. The British Cardiovascular Intervention Society used a plot of observed outcome against the confidence interval of predicted outcomes in their 2015 public reports of operator and hospital outcomes (Fig 3.8; www.BCIS.org.uk). The display of data becomes more complex where there is a desire to show comparative rankings of healthcare providers. ‘League tables’ and ‘caterpillar plots’ have been used, but can be misleading. In Fig 3.9, it appears that as one moves from left to right along the horizontal axis, successive providers demonstrate worsening outcomes. At least this figure does include error bars and so a judgement can be made regarding the possible statistical significance of the differences. Nevertheless, the focus is on a spurious ranked order of results (66).

33

34

34

Section 1  

background and basics

Box 3.1  Definitions used during the calculation of the SYNTAX score The SYNTAX score is calculated by a computer program consisting of sequential and interactive self-​guided questions. All the below mentioned definitions are projected in a side window when the signal (i)  indicating information, available for each question, is pointed with the cursor.

Box 3.2  The SYNTAX score algorithm 1. Dominance 2. Number of lesions 3. Segments involved per lesion Lesion characteristics 4. Total occlusion

Definitions

i. Number of segments involved

Dominance. a) Right dominance: the posterior descending coronary artery is a branch of the right coronary artery (segment 4). b) Left dominance: the posterior descending artery is a branch of the left coronary artery (segment 15). Co-​dominance does not exist as an option in the SYNTAX score. Total occlusion TIMI 0 flow: no perfusion; no antegrade flow beyond the point of occlusion Bridging collaterals Small channels running in parallel to the vessel and connecting proximal vessel to distal and being responsible for the ipsilateral collateralisation Trifurcation A  junction of three branches, one main vessel and two side-​branches. Trifurcations are only scored for the following segment junctions: 3/​4/​16/​16a, 5/​6/​11/​12, 11/​12a/​12b/​13, 6/​7/​9/​9a, and 7/​8/​10/​10a Bifurcation A junction of a main vessel and a side branch of at least 1.5 mm in diameter. Bifurcations are only scored for the following segment junctions: 5/​6/​11, 6/​7/​9, 7/​8/​10, 11/​13/​12a, 13/​14/​14a, 3/​4/​16, and 13/​14/​15. Bifurcation lesions may involve one segment (types A, B, and E); two segments (types C, F, and G), or three segments (type D) Aorto ostial A lesion is classified as aorto-​ostial when it is located immediately at the origin of the coronary vessels from the aorta (applies only to segments 1 and 5, or to 6 and 11 in case of double ostium of the LCA) Severe tortuosity One or more bends of 90° or more, or three or more bends of 45° to 90° proximal of the diseased segment Length >20 mm Estimation of the length of that portion of the stenosis that has a ≥50% reduction in luminal diameter in the projection where the lesion appears to be the longest. (In the case of a bifurcation lesion at least one of the branches has a lesion length of ≥20 mm) Heavy calcification Multiple persisting opacifications of the coronary wall visible in more than one projection surrounding the complete lumen of the coronary artery at the site of the lesion. Thrombus Spheric, ovoid, or irregular filling defect or lucency surrounded on three sides by contrast medium seen just distal or within the coronary stenosis in multiple projections or a visible embolization of intraluminal material downstream. Diffuse disease/​small vessels More than 75% of the length of the segment has a vessel diameter of 2 mm, irrespective of the presence or absence of a lesion. LCA, Left coronary artery; TIMI, Thrombolysis in Myocardial Infarction Study Group.

ii. Age of the total occlusion (>3 months)

Reprinted from Sianos G, Morel M-A, Kappetein AP et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease EuroInterv. 2005;2:219–27, with permission from Europa Edition.

iii. Blunt stump iv. Bridging collaterals v. First segment beyond the occlusion visible by antegrade or retrograde filling vi. Side branch involvement 5. Trifurcation i. Number of segments diseased 6. Bifurcation i. Type ii. Angulation between the distal main vessel and the side branch 20 mm 10. Heavy calcification 11. Thrombus 12. Diffuse disease/​small vessels i. Number of segments with diffuse disease/​small vessels Reprinted from Sianos G, Morel M-A, Kappetein AP et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease EuroInterv. 2005;2:219–27, with permission from Europa Edition.

This use of confidence intervals to differentiate between a significant and insignificant difference will always identify 5% as outliers and, for additional reasons, has been shown to be rarely appropriate (67). A  control chart constructed from the same information provides a much clearer representation of the data (Fig 3.10). There is no ranking of providers, and the few outliers are easy to identify (marked 19, 32, 35). Hospital 19 is much more obviously an outlier than can be appreciated from the league table. However, the control limits on this chart do not take account of the fact that confidence intervals surrounding low volume activity will be wider than around one with high throughput. This brings us to funnel plots described in some detail by Spiegelhalter (68). A funnel plot charts the observed rate of an event against the volume of activity. Superimposed are the 95% (approximately 2 SDs) and 99.8% (approximately 3 SDs) prediction limits and the overall mean event rate. The plot is funnel-​shaped because the random variation in outcomes is higher if fewer procedures are performed. Conversely, in hospitals with high volumes, the confidence

35

Chapter 3 

intervals are tighter and the degree of certainty about the observed outcomes is greater. These plots are a very useful way to depict large amounts of data graphically, and easily highlight the variations in confidence ranges of the point estimates that exist between different institutions with different levels of activity. To demonstrate the benefits of this form of display see Fig 3.11, showing the data for in-​hospital mortality following CABG in New York State in 1997–​99 (69). The observed mortality has been divided by expected mortality of each unit to find the standardized mortality rate. This has been multiplied by the overall state-​wide mortality rate (2.2% in 1997–​99) to obtain a risk-​adjusted mortality rate. There is no particularly obvious way to analyse these data visually when presented in this format. Fig 3.12 is the equivalent funnel plot. From the latter representation, it is clear that there is a high-​ mortality, low-​volume hospital as well as a borderline high-​volume, low-​mortality unit. Most units sit well within the funnel, their spurious ‘ranking’ not being a strong visual signal. The plots also lend themselves to assessment of the relationship between volume A

risk assessment and analysis of outcomes

and outcome, and there is suggestion of a trend towards lower mortality in higher volume units. Greater variability in measured outcomes within lower volume centres is allowed. The Society for Cardiothoracic Surgery in Great Britain & Ireland used a version of these plots to display the outcomes of individual operators and hospitals in their 2015 public outcomes report (www. scts.org). The funnel shape is derived from the analysis of the entire dataset of outcomes of all operators and provided as the background, but only the individual observation in question is marked (Fig 3.13). This allows an appreciation of the context of that single analysis without the risk that inappropriate comparisons between centres or operators are made.

Limitations in methods used to detect outlying performance Whatever risk adjustment model is used, and however the results are displayed, the aim is to try to identify outcomes that are significantly different from what would be expected—​that is, to identify outliers. Commonly, cut points are used to define outliers if they Lesion 1 Segment 5: 5x2 + Bifurcation Type A + Heavy calcification Lesion 1 score:

10 1 2 13

Lesion 2 Segment 6: 3.5x2 + Bifurcation Type A + Angulation 50%

LAD > 50%

Lesion 3 Segment 11: 1.5x5 Age T.O. is unknown + Blunt stump + side branch First segment visualized by contrast: 13 + Heavy calcification + Length Lesion 3 score: 14.5

LCX 100%

RCA 100% SYNTAX SCORE 54.5

Lesion 4 Segment 1: 1x5 Age T.O. is unknown + Blunt stump + side branch first segment visualized by contrast: 4 + Tortuosity + heavy calcification + Length core:

7,5 1 1 1 1 2 1

5 1 1 1 3 2 2 1 16

Figure 3.6  Two examples demonstrating the calculation of the SYNTAX score. LAD, Left anterior descending (coronary artery); LCX, left circumflex coronary artery; LM, left main stem; RCA, right coronary artery; t.o., total occlusion. Reproduced from Sianos G, Morel M-A, Kappetein AP, et al. The SYNTAX score: an angiographic tool grading the complexity of coronary artery disease. EuroInterv 2005; 1:219–27, with permission from Europa Edition.

35

36

36

Section 1  

background and basics B

Lesion 1 Segment 6: 3.5x2 Lesion 1 score:

7 7

Lesion 2 Segment 11: 1.5x2 + Tortuosity Lesion 2 score:

3 2 3

Lesion 3 Segment 1: 1x2 Lesion 3 score:

2 2

Lesion 4 Segment 1: 1x2 + tortuosity + Length Lesion 4 score:

2 2 1 5

LAD > 50%

LCX > 50%

RCA2 > 50%

RCA3 > 50%

SYNTAX SCORE 17 Figure 3.6  (Continued)

fall outside either 2 or 3 SDs from the mean, and these cut-​offs have been labelled ‘alert’ (2 SDs), and ‘alarm’ (3 SDs). Although this process is at the heart of clinical audit and the question being asked is clear, the statistical methodology is fraught with complexity, and there is a need to achieve the correct balance. If an outlier is defined only when performance is markedly different from predicted, then patients are inadequately protected. Yet taking lower cut-​off levels entails the risk of incorrectly criticizing perfectly satisfactory performance. Understanding some of the potential limitations of the methods used will help to guide an appropriate balance and provide perspective when assessing an individual’s or a hospital’s performance. All risk-​adjusted methods are dependent on the appropriate calibration and sensitivity of the risk adjustment model, as discussed above inn the section on ‘Risk adjustment models: creation and assessment’, and the accuracy of the data recorded for the cases being analysed. In addition, the mathematics behind the application of models to the assessment of outlier status are complex, and several different statistical methods can be used. Frequently the exact binomial distribution is used to identify outliers. If overdispersion is identified (using a Kolmogorov–​Smirnov test, for example) then adjustments can be applied to rescale the outcome P values (Winsorization). An

alternative approach is to use random effects (multilevel/​clustered) analyses, which may have advantages in reducing the number of assumptions that need to be made. Each method has benefits and limitations, and different methods can lead to the identification of different outliers. This is an area of statistical methodology that is still evolving, and no agreed gold standard technique yet exists. One of the most important issues is how best to adjust for multiple comparisons. A  number of statistical tools have been used, including, for example, the Bonferroni correction, although this has been criticized as being too generous. The ‘false discovery rate’ provides an intuitive way to grasp the concept of multiple comparisons. If we are assessing a single operator, and that operator is performing as would be expected by the risk adjustment model, then there is only a 2 in 1000 (0.2%) chance they will fall outside a 3-​SD cut-​off. This means they will incorrectly appear to perform less well than expected 1 in 1000 times; that is, the false alarm rate is only 0.1%. But what happens if we make this assessment with multiple operators, and at multiple times? If we consider 100 operators, this means that each year the chance of incorrectly identifying an outlier at 3 SDs is 0.1% of 100 = 0.1, so that this will occur every 10 years (reciprocal of 0.1). However, as the number of comparators

37

Chapter 3 

risk assessment and analysis of outcomes

100

4-year mortality (%)

80

60

58.4%

40 31.5% 20

15.1%

0

3%

1.3%

0.6%

6.8%

20

0

40

60

80

100

Total points

Points SYNTAX score

CABG 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

PCI 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

0+

0

10 20 30 40 50 60

Age (years) 50

40 CrCl (mL/min)

LVEF (%)

90

60

60

70

80

40 50 60 70 80 90

30

50 40 30 20 1

50

0

1

60

30

40

30

20

0

Left main

Sex

M

F

F

M 1

1

COPD 0

0

1

1

PVD

0

0

Figure 3.7  Nomogram for SYNTAX Score II. CABG, Coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; CrCl, creatinine clearance; LVEF, left ventricular ejection fraction; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease. Reproduced from Farooq V et al. Anatomical and clinical characteristics to guide decision making between coronary artery bypass surgery and percutaneous coronary intervention for individual patients: development and validation of SYNTAX score II. Lancet 2013; 381: 639–​50 with permission from Elsevier.

rises so does the chance of a false alarm. If we assess 600 operators, a false alarm would occur every 2 years. Thus, as we make multiple comparisons at multiple time points the interpretation becomes more complex. The statistical certainty that an outlier is correctly identified is not 99.9%; rather it is dependent on the total number of operators being assessed and the total that appear to be outliers. The concept of a ‘false discovery rate’ (also known as controlling for type 1 errors) was first proposed

in the mid 1990s (70) and has since been used to deal with the statistical problems that arise from multiple comparisons (71). The calculation of the ‘false discovery rate’ is given in the equations below: ◆ False ◆

discovery rate = E/​k

Probability that an identified outlier is actually an outlier = 1–​(E/​k) E is the expected number of false alarms k is the number of outliers actually observed.

37

38

Section 1  

background and basics they lie away from the cut-​off point. Other factors therefore need to be taken into account to give face validity to statistical assessments such as the actual and predicted rates of adverse outcomes. A sensitivity analysis can be performed to see how the assessment changes with a single extra case with good or poor outcome, and longitudinal measures to see if outlying performance is isolated or repeated over several time periods (recognizing the characteristic of regression to the mean). Ultimately these methods can only be used as a guide to finding potentially poor practice. In the end it is important to try to make a balanced judgement that seeks to protect the rights of both operators and patients. An alert or alarm, however defined, should trigger appropriate individual and organizational scrutiny to try to identify potential issues and try to improve the quality of care provided.

30-Day Post-PCI Survival (ONS Tracked):

National average = 97.98% 94%

96%

98%

100%

Percentage of Tracked Cases Surviving to 30 Days Post–Procedure Percentage of Tracked Cases Surviving to 30 Days 99.8% Control Limits for Predicted 30-day Survival

Sequential assessment

Figure 3.8  Example of operator data from the British Cardiovascular Intervention Society’s public reports published online in 2015. The horizontal axis shows survival, running up to 100% at the far right. The green bar shows the range of survival that would have been expected given this operator’s case mix. The black dot represents that observed outcome. ONS, Office of National Statistics; PCI, percutaneous coronary intervention.

Aside from difficulties in analysis, there are several disadvantages with cross-​sectional assessments of performance. The findings are always out of date; for example, the New York State PCI data are not available until 3 years after collection (the 2010–​2012 data were available in October 2015 [72]), and for the 2015 public reports on outcomes in the UK the PCI data was for procedures performed in 2012, 2013, and 2014, and the cardiothoracic surgical data were of operations performed from April 2011 to March 2014. It is also difficult to detect gradually changing trends, and yet it is only by doing this that there is the hope of detecting a move towards outlier performance, and so reacting to potential problems before true outlier status is actually reached and any harm done. This is where sequential or case-​ by-​ case monitoring techniques have a pivotal role. A  variety of different graphical and statistical process control methodologies have been developed. These methods date back to the 1920s when Shewhart developed graphical displays to try to improve the quality of telephones being manufactured in Bell laboratories (7)  and in the case of cumulative sum charts, to clinical chemistry laboratories of the 1950s (41, 73).

Figure produced from the public reports prepared by National Institute of Cardiovascular Outcomes Research (NICOR) on behalf of BCIS.

So, if there were 1000 cardiologists being assessed, we would expect 1 false alarm at the 3-​SD level (1 × 0.001). If 5 were found to be outliers, then there is a 20% chance that the observation is incorrect, and that satisfactory practice is being incorrectly labelled as substandard, and an 80% chance that they are correctly identified as truly performing below an expected level. The examples above underscore the point that we need to look at more than a single metric in trying to identify poor performance correctly. Although dichotomous cut points are used (2 SDs or 3 SDs), it is obvious that there is a progressively increasing statistical likelihood that an outlier is correctly identified the further

25

20 30-day mortality (%)

38

15

10

5

0

32

7

8 2

16

10

3 34

12

5 25

33

22 11

30

28 17

13

26 31

23

29 15 36 27 18 6 37 20 14 Overall 4 1 21 9 19 35 24

Hospital ID No

Figure 3.9  League table presentation of outcome data. Reproduced from Adab P, Rouse AM, Mohammed MA, et al. Performance league tables: the NHS deserves better. BMJ 2002; 324 (7329):95–​8, with permission from BMJ Publishing Group Ltd.

39

Chapter 3 

risk assessment and analysis of outcomes

No. of deaths within 30 days

100 90 80

19

70 60 50 40 30

35

24

20 32

10 0

200

300 400 No of patients admitted at each hospital

500

600

Expected mortality (11.2%, based on combined death rate of all hospitals) Lower control limit Upper control limit Observed hospital mortality

Figure 3.10  Control chart presentation. Reproduced from Adab P, Rouse AM, Mohammed MA, et al. Performance league tables: the NHS deserves better. BMJ 2002; 324 (7329):95–​8, with permission from BMJ Publishing Group Ltd.

A key principle behind sequential monitoring charts is intuitive interpretation of the data. However, avoiding misinterpretation requires an understanding of the way they are produced and therefore the assumptions on which they rest. The original Shewhart chart is shown in Fig 3.14. This shows that if the outcome of interest is ‘in control’, not deviating from that expected, it marches along the graph in a horizontal line. Random variation will allow this line to deviate up or down, while keeping out of the areas of alarm. Warning limits suggest the system may be out of control, and if the line crosses the upper or lower alarm limit, then there is a high statistical likelihood that the system is out of control and special cause variation has been identified. Almost all the charts described below follow the same general principle of a cumulative ‘observed–​ expected’ plot with horizontal thresholds.

Cumulative sum The Shewhart control charts were designed to monitor batches of results. Cumulative sum (CUSUM) charts can be used to assess sequential individual procedures, and were the first group of charting methods widely applied to the assessment of healthcare outcomes. For the following examples let us assume that a procedure is performed, and that the result is either successful or unsuccessful. In its most basic form, the CUSUM chart simply plots success or failure for each successive procedure from zero. Success causes the line to move horizontally, failure causes the line to move up one step (Fig 3.15). It is possible to construct boundaries for this chart, based on the expected probability that any procedure will result in success or failure, but this makes the assumption that the probabilities are the same for every procedure. This method was used very effectively in the assessment of neonates undergoing an arterial switch operation (74). For a more detailed explanation and mathematical derivations see Rogers (75).

Risk-​adjusted modifications of cumulative sum When attempting to tease out modifiable factors in patient outcome we must account for expected variation based on risk prediction models, and determine whether observed performance is as would be predicted. To this end, there have been a number of modifications to the fundamental CUSUM chart. One of the earliest described is the variable life-​adjusted display (VLAD) chart (76), also called cumulative risk-​adjusted mortality (CRAM) (77) and risk-​ adjusted sequential probability ratio test (SPRT) (78) charts. The VLAD or CRAM plot is constructed as follows. The graph starts at zero. A successful outcome causes the line to increment, and an unsuccessful outcome causes the line to decrement. The amount the line rises or falls with each case depends on the predicted risk of an adverse outcome before the procedure took place. The calculation is quite simple. If the probability of an adverse outcome (as predicated by the risk model for that particular procedure) is P, then after a case with a successful outcome, then line rises by P. If there was an adverse outcome, then the line falls by 1 P. Thus, if a high-​risk case is carried out without a complication, then a lot of credit is accrued and the line rises more than if a low-​ risk case is undertaken without complication. Conversely, if a low-​ risk case has an adverse outcome, then the line falls more than if the adverse outcome occurs in a high-​risk case. Overall, after a sequence of cases, if the observed outcomes are similar to the outcomes that would have been predicted by the model, then the line will tend to be horizontal. More failures than predicted by  the model will cause the line to fall, and more success will cause it to rise (Fig 3.16). Interpretation of the display is straightforward but one disadvantage is difficulty in the construction of control limits. Formal statistical methods for sequential analysis have been developed and are used in risk-​adjusted SPRT charts (79–​81). These charts use a running log-​likelihood ratio, increased or decreased

39

40

40

Section 1  

background and basics Albany Medical Center Arnot-Ogden Bellevue Beth Israel Buffalo General Columbia Presbyterian-NYP Ellis Hospital Erie Country LIJ Medical Center Lenox Hill Maimonides Millard Fillmore Montefiore-Einstein Montefiore-Moses Mount Sinai New York Hospital-Queens NYU Hospitals Center North Shore Rochester General St. Elizabeth St. Francis** St. Josephs St. Lukes-Roosevelt St. Peters** St. Vincents Strong Memorial United Health Services Univ. Hosp.-Stony Brook Univ. Hosp.-Upstate Univ. Hosp. of Brooklyn* Weill Cornell-NYP Westchester Medical Center Winthrop Univ. Hosp.

Key Risk-adjusted mortality rate Potential margin of statistical error * Risk-adjusted mortality rate significantly higher than statewide rate based on 95 percent confidence interval. ** Risk-adjusted mortality rate significantly lower than statewide rate based on 95 percent confidence interval.

0

2 4 2.24 New York State Average

6

8

10

12

14

Figure 3.11  Plot of in-​hospital mortality following coronary artery bypass graft in New York State in 1997–​99.

after each observation by an amount dependent on the observed outcome, and that predicted from the risk model. The risk-​adjusted SPRT charts have the advantage of providing a formal statistical test of significance, but the chart has a less transparent interpretation. There have been other descriptions of risk adjustment for CUSUM, for example by Steiner et al. (82).

Cumulative funnel plots A method that combines sequential assessment and the advantages of funnel plots has been demonstrated in patients being treated by PCI (83). Its main advantage over CUSUM styled charts is that, while data on a CUSUM chart combines observed and predicted outcomes, cumulative funnel plots show observed and expected outcomes as separate lines. This provides more visual clues to help interpret the findings. The case mix becomes more obvious (evidenced by whether the predicted adverse event rate is high or low), and the way this compares to the observed outcomes is readily apparent. As the number of cases in the sequence rises, the influence of random fluctuation falls and the confidence intervals narrow. To construct these plots the cumulative mean predicted adverse event

rate is calculated as each successive case is added to the series, and plotted as a line, with another line drawn from the mean cumulative actual adverse event rate. (Figs 3.17, 3.18, and 3.19) An example of real data from a PCI centre is provided in Fig 3.20, to show a comparison with a VLAD plot of the same data. The VLAD plot shows that, for the first 29 cases, there are no adverse outcomes, and the plot rises as observed outcome exceeds that predicted by the model. Then there is a complication and the plot steps down, crossing the horizontal axis and so showing worse outcomes than predicted by the model. The plot then moves up gradually until it once again cross above the x axis, outcomes better than model prediction. It steps down again with a complication for case 86. At case 118 it steps up suddenly as a high-​risk case is undertaken without a complication. The third downward step is the result of the third adverse event in this series at procedure 127. As the plot continues, there is an overall downward trend, suggesting that the outcomes are progressively worse than the model would predict. Compare this with the cumulative plot of the same data. The observed and predicted information is no longer combined into one line, but separated into observed data (blue line) and predicted (red

41

Chapter 3 

risk assessment and analysis of outcomes

% mortality

8 6 4 2

0 0

1000 Volume of cases

500

99.8 % limits

1500

2000

99 % limits

Figure 3.12  Speigelhalter’s funnel plot of in-​hospital mortality following coronary artery bypass graft in New York State in 1997–​99.

line). For the first 29 procedures there are no complications and the mean observed rate runs along at 0%. The first adverse event lifts the line to 3.4% (1 case in 29), and it crosses the predicted line. The mean cumulative observed rates fall again, re-​crossing below the predicted line until the second adverse event, so creating the sawtooth pattern. The running mean of the predicted rates (red line) gives a backdrop against which observed rates can be judged. The jump up in predicted event rate can be seen after high-​risk case 118, as in the VLAD. As the series progresses, and the means are calculated from an ever larger number of procedures, the variations in the lines become less marked (unlike in the corresponding VLAD plot). It can also be appreciated that, by the end of this sequence, the outcomes are poorer than the model would have predicted (the observed mean is about 2.5% and the predicted about 1.8%). Unlike

the VLAD plot we can see these actual values using the cumulative method. To then gauge whether the observed difference is due to random variation, confidence intervals are added to create the final part of the funnel plot (Fig 3.21). These charts provide a visually intuitive and informative way to evaluate and communicate sequential data, and have been adopted by the British Cardiovascular Intervention Society. Every 3 months a cumulative funnel plot of each PCI centre’s risk-​adjusted outcomes is sent to each UK centre to help inform their audit and quality improvement programmes.

Statistical process control charts for process Some aspects of treatment by PCI can be assessed by simply looking at the speed of treatment. For emergency PCI in the treatment of STEMI, delays to provision of PCI must be minimized. Every step

Data For Period April 2011–March 2014 Risk-adjusted In-Hospital Survival Rate

100.0%

Risk-Adjusted Survival Rate

97.5%

95.0%

509 operations with a risk-adjusted survival rate of 96.93%

92.5%

250

500 Number of Operations Predicted Value

750

1000

99.8% Control Limits

Figure 3.13  Funnel 99.8% confidence limits surrounding the risk-​adjusted outcomes of a single surgeon, from the public reports of surgical outcomes from the Society for Cardiothoracic Surgery in Great Britain & Ireland, published online in 2015. Graph courtesy of the National Adult Cardiac Surgery Audit, managed by the National Institute of Cardiovascular Outcomes Research (NICOR) and commissioned by the Healthcare Quality Improvement Partnership (HQIP).

41

42

Section 1  

background and basics Low probability event

Outcome of interest

Upper warning limit Upper alarm limit Expected variation Centreline

Lower alarm limit Lower warning limit Low probability event

0

10

20

30 40 50 Patient number

60

70

80

Figure 3.14  Control chart by Shewhart.

in the pathway, from a patient’s call for help to the opening of an occluded coronary artery, must be as swift as possible. A variety of methods have been used to try to describe these delays, with the intention of minimizing them. A common technique is to describe the percentage of patients treated within a certain target, such as door-​to-​balloon times of less than 90 minutes. Performance can be compared either between different healthcare providers or within any one provider for different time periods. The resulting statistics are relatively insensitive to the occasional outlier patient in the same way to that of a median value. However, these outliers need to be identified and assessed to find out if there was a preventable problem with the process of care. Such evaluation is facilitated by data display using statistical process control charts. For an example see Figure 1 from Hall et al. (84). Successive cases are plotted on the x axis with the measured time period on the y axis. Median values for the time points of interest are marked as a horizontal line with confidence intervals added to provide an upper control limit, a statistically defined boundary roughly equivalent to the 99% confidence limit. Individuals with prolonged delays sit well above the control line and so become evident. Overall changes to median times can also be observed.

Unresolved issues regarding statistical process control methods

1 0.8 0.6 0.4 Better 0.2 Performance 0 −0.2 Worse −0.4 −0.6 −0.8 −1

5 4 3 2

Case 1

Case 2

Case 3

Case 4

built up over a sequence of procedures, then a change to subsequent poorer outcomes will take longer to cross a warning boundary. Yet a shorter series will have wide confidence intervals and little statistical power to detect outliers. There is also the sense that these charts are attempting to provide contemporaneous measurements, and so should reflect current rather than past performance. Rather than taking an arbitrary cut-​off, an exponential memory loss called ‘exponentially weighted moving average’ has been proposed (74). It has also been suggested that SRPT charts are reset whenever an individual crosses the lower boundary line of ‘performance as expected’, as at that stage monitoring is no longer of interest and it might be appropriate to start a new monitoring period to retain sensitivity to changes in performance line (78). This approach has been used in the risk-​adjusted CUSUM (82). While similar to the SPRT chart, it is constrained to lie above zero, which means that it cannot build up credit, and thus retains sensitivity to underperformance. However, the more times the line is restarted, the greater the likelihood of a type 1 statistical error (i.e. rejecting the hypothesis that performance is satisfactory even though it is). Indeed, this error becomes a certainty after an infinite number of restarts. Another approach is to try to decide on the ‘average run length’ for a period of sequential monitoring based on the average number of observations needed

50 45 40 35 30 25 20 15 10 5 0

Observed cumulative mean event rates 25%

11%

20%

Ca se 0 Ca se 1 Ca se 2 Ca se Ca 3 se 4 Ca se 5 Ca se 6 Ca se 7 Ca se 8 Ca se Ca 9 se 1 Ca 0 se 11

1

Start

Figure 3.16  A variable life-​adjusted display plot. Case 1: the risk model predicts the probability of a failed outcome as 5% and the case is successful, the line rises by 0.05. Case 2: the predicted risk of failure is 60% and the procedure was again successful, the line rises more steeply, by 0.6. Case 3: predicted risk here was 70% and there was an adverse outcome, the line falls by 0.3 (1–​0.7). Case 4: predicted risk of 5% was also a failure; the line falls more steeply, by 0.95.

Cumulative adverse event rate %

CUSUM techniques have no memory loss. A decision needs to be made about when and how to discard past experience—​in other words when to reset the chart to zero. If a large amount of ‘credit’ is

Cumulative failures

42

0 0

1

2

3

4

5

6

7

8

9

10

Procedure number

Figure 3.15  A cumulative sum chart showing success following the first two procedures, then failure. Procedure 3 therefore causes line to move up one unit.

Figure 3.17  The plot of actual outcomes: the first three procedures occur without an adverse event (running mean event rate = 0%). Case 4 is complicated by an adverse event, lifting the running mean from 0 to 25%. The next few cases are uncomplicated so that by case 9 there has only been one adverse event in nine cases. The running mean has drifted down to 1 in 9, or 11%. Case 10 has a complication, and the running mean rises to two in ten, or 20%.

43

50 45 40 35 30 25 20 15 10 5 0

10% 5%

Predicted MACCE

Running Mean

Figure 3.18  For each case, the adverse event rate predicted from the risk model used is calculated. A running mean is then plotted. If the first two cases had a predicted risk of 10% and 5%, respectively, then the mean predicted risk by case 2 is 7.5% and so on. Exact binomial method data sourced from Wu et al. (2006) A risk score to predict in-​hospital mortality for percutaneous coronary interventions. J. Am. Coll. Cardiology (47)3:654–​660.

before a conclusion can be drawn. The calculations of these run lengths is not straightforward, and will depend on the risk profile being studied. If this changes, then the calculated average run time may no longer be appropriate. Such calculations can be made using the Markov chain procedure (82, 85). There is no current consensus as to the best way to handle the issue of sequential chart memory. These techniques have tremendous value in the assessment and improvement of healthcare delivery. All have their advantages and disadvantages. CUSUM-​based methods are very sensitive to subtle shifts away from predicted outcomes but, if they prove too sensitive, they will create false alarms and will be discredited. Equipoise must be achieved between systems that are sensitive enough to warn of clinically important changes in outcomes but do not trigger alarms where no clinically important shift exists. Interpretation of these charts must account for their limitations. They should be a guide to monitoring performance which primarily display data, not assess its true statistical significance. Observed and predicted cumulative mean event rates

50 45 40

Event rate %

35 30 25 20 15 10 5 0 0

1

2

3

4

5

6

7

8

9

10

Case number

Observed

risk assessment and analysis of outcomes

Problems with publication of outcome data

Predicted cumulative adverse event rate

Ca se 0 Ca se 1 Ca se 2 Ca se Ca 3 se 4 Ca se 5 Ca se 6 Ca se 7 Ca se 8 Ca se Ca 9 se 1 Ca 0 se 11

Cumulative adverse event rate %

Chapter 3 

Predicted

Figure 3.19  These two running means are then plotted on the same chart and can be compared.

Our ultimate goal in championing these methods is the provision of optimal patient care. We need to measure how well we are treating patients to develop strategies to improve this care. Purely clinical components cannot, however, be isolated from political and social factors. The act of reporting data improves a unit’s own scrutiny of its performance and increases attention to processes, appropriateness, and quality of care (the Hawthorne effect). Public reporting is intended to nurture a culture of openness and accountability to guarantee quality assurance for patients, commissioners, and regulators. Yet paradoxically there are compelling data that such a mechanism can become deleterious to patient care. Reported advances in outcomes may be due to genuine improvements in patient management, but important factors can suggest apparent improvements where none has taken place. Such illusions occur with alterations in patient selection and manipulation of the risk model. Both have occurred since public reporting and neither is beneficial.

Risk-​averse behaviour The change in patient selection is the most disquieting, as it means that the sickest patients at highest risk of adverse outcomes, but with potentially greatest gains from intervention, are turned down for treatment. This is described as ‘risk-​averse’ behaviour. It is reinforced by difficulties in developing models that adequately account for the severity of illness in extremely sick patients, making them poor at differentiating risk levels within the high-​risk cohort. The most studied example of this effect followed the public reporting of outcome for CABG procedures in New  York State. Surgeons became reluctant to operate on the sickest patients (86), there was an increase in the number of high-​risk patients transferred out of state for their surgery (87), and patients from ethnic minorities were less likely to be offered surgery (88). Across the UK the public reporting of CABG outcome has also altered patient selection, with surgeons exhibiting risk-​averse practice. There have been attempts to measure this objectively (89), but the analyses are fraught with difficulty because national datasets do not record those patients for whom surgery was not offered (90). Similar behaviour has been observed with PCI. Public reporting seems to be the most compelling explanation for the big differences in case mix and in-​hospital mortality between two large quality-​ controlled regional PCI registries: Michigan (no public reporting) and New  York (public reporting) (91). In Massachusetts, public reporting of PCI outcomes was accompanied by a 43% fall in the number of patients being treated with cardiogenic shock (2.28% in 2003, compared to 1.2% in 2005) (92, 93). In addition, it has been demonstrated that, once a hospital was publicly reported as being an outlier, the patients subsequently treated at that institution were of significantly lower predicted risk in subsequent years, implying risk-​averse behaviour. Furthermore, there was no associated increase in predicted risk at non-​outlier institutions, which might suggest that sicker patients were being transferred for treatment elsewhere (94). A retrospective analysis of the SHOCK trial is in line with these conclusions. Patients presenting in New  York with cardiogenic shock were less likely than those presenting in other regions to be treated by PCI or CABG. After propensity matching, the odds ratio for being treated by PCI was 0.51 (95% confidence interval 0.33–​0.77, P = 0.002). While there was no significant difference in

43

4

Section 1  

background and basics VLAD Display

Worse

Cumulative Risk adjusted difference predicted - observed MACCE

0.5 0

50

100

150

200

250

300

−0.5 −1 −1.5

Case 29

Case 86 Case 118

−2 −2.5 −3

Sequential PCI procedures Cumulative Funnel

4.0% 3.5% 3.0% Worse

2.5% Risk

44

2.0% 1.5% 1.0% Case 118

0.5% 0.0%

0

50

100 150 200 250 Sequential PCI procedures performed Observed MACCE

300

Predicted MACCE

Figure 3.20  Variable life-​adjusted display and cumulative funnel plots of the same data from a percutaneous coronary intervention unit. MACCE, Major adverse cardiac and cerebrovascular events; PCI, percutaneous coronary intervention; VLAD, variable life-​adjusted display.

mortality for patients who did receive revascularization, those untreated had a 1.5-​fold higher mortality rate (95). Assessing treatment and outcome of all patients presenting with a clinical syndrome, rather than just those who are treated by PCI, provides a more overarching insight. In a comparison of states with public reporting (New  York and Massachusetts) to those without (Connecticut, Maine, Maryland, New Hampshire, Rhode Island, and Vermont), the treatment and outcomes of all patients presenting with acute myocardial infarction was assessed. Public reporting was associated with a lower use of PCI for treatment, particularly in patients at higher risk (older patients, presentation in shock or with STEMI). Not surprisingly, the outcomes following PCI in the selected lower risk cohort were better, but the outcome of those not being treated by PCI was worse, and the overall in-​hospital mortality of all patients was higher in states with public reporting (96). Resnic and Welt describe a useful framework map of relative risks and benefits, and use this to demonstrate the case selection ‘creep’ that occurs as high-​risk patients are left untreated (Figs 3.22 and 3.23) (97).

For PCI, in addition to a concern that risk models may not adequately adjust for the highest risk cases, there is the additional concern that a large proportion of the mortality following a PCI is not actually related to the PCI procedure. Thus when mortality occurs after a PCI, it occurs in spite of, not because of, the procedure. In a study into the cause of death after 4078 PCIs at the Cleveland Clinic between 2009 and 2011, only 58% died of cardiac causes at 30 days, and 42% were related to the PCI procedure. The commonest non-​cardiac cause of death was sepsis, and the next most frequent neurological, of which withdrawal of care after anoxic brain injury was the most prevalent (98).

‘Gaming’ There is robust evidence that hospitals have manipulated the predictions of risk models, termed ‘gaming’. Following the introduction of the Cardiac Surgery Reporting System in New York, there was a large and sudden increase in the reported prevalence of five key risk factors (renal failure, congestive heart failure, chronic

45

Chapter 3 

Cumulative funnel with statistical limits inluded

20%

Likelihood of survival from procedure

18% 16% 14% Risk

12% 10% 8% 6% 4% 2% 0%

Minimal CAD w/o Ischemia Focal CAD w/o Angina Focal CAD w Angina

Case selection “creep”

Unstable Angina 55yo with NSTEMI 65yo with Ant. STEMI 50yo STEMI in Shock 75yo ST ∆’s w/Sepsis

0

50

100 150 200 250 Sequential PCI procedures performed Observed MACCE 3 SD

300

Predicted MACCE 2 SD

Figure 3.21  The final cumulative funnel plot. Using an exact binomial method (68), lines are added that delineate the 95% and 99.8% prediction limits. As the number of cases in the sequence rises, so the confidence intervals narrow, giving rise to the characteristic funnel shape of the prediction limit boundaries. The scale of the axis needs to be modified to accommodate the control limits added and it is clear that the small differences observed are within limits of random fluctuation. Evident also in this sequence of 280 patients is that variation in statistically accepted outcome even at 2 SDs is very large, at between approximately 0 and 4%. MACCE, Major adverse cardiac and cerebrovascular events; PCI, percutaneous coronary intervention.

obstructive pulmonary disease [COPD], unstable angina, and low ejection fraction). Thus from 1989 to 1991 reported COPD increased from 6.9% to 17.4%, congestive heart failure from 1.7% to 7.6%, and renal failure from 0.4% to 2.8%. Thess increases in reported prevalence of 152%, 347%, and 600% could not be explained by a genuine change in patient demographics (99). The

Likelihood of survival from procedure

risk assessment and analysis of outcomes

Low risk Low benefit

Low risk, High benefit

High risk Low benefit

High risk, High benefit

Incremental patient health benefit

Figure 3.22  Map of percutaneous coronary intervention risk versus clinical benefit. The vertical axis denotes the risk of the procedure represented as the likelihood of survival to hospital discharge. The horizontal axis denotes the patient benefit represented as the incremental health benefit of having the procedure performed. Reprinted from JACC, 53(10), Resnic FS and Welt FGP, The public health hazards of risk avoidance associated with public reporting of risk-​adjusted outcomes in coronary intervention, 825–​30, (2009) with permission from American College of Cardiology Foundation, published by Elsevier.

75yo STEMI in Shock

Incremental patient health benefit

Figure 3.23  Potential for ‘risk avoidance creep’. The map of percutaneous coronary intervention risk versus clinical benefit (see Fig 3.22) is shown with illustrative examples. Green ovals indicate scenarios in which clinical benefit is high; grey ovals indicate intermediate risk; and red ovals indicate scenarios in which incremental risk is negligible. The red dashed arrow indicates the ‘risk avoidance creep’ toward lower risk cases in the face of public reporting. CAD, Coronary artery disease; NSTEMI, non-​ST-​segment elevation myocardial infarction; STEMI, ST-​segment elevation myocardial infarction; w, with; w/​o, without; yo, years old. Reprinted from JACC, 53(10), Resnic FS and Welt FGP, The public health hazards of risk avoidance associated with public reporting of risk-​adjusted outcomes in coronary intervention, 825–​30, (2009) with permission from American College of Cardiology Foundation, published by Elsevier.

effect was to artificially elevate predicted risk, making observed outcomes relatively better than predicted, even if there had been no actual change in the quality of the care. Canadian investigators reported that, in Ontario, mortality associated with CABG was sharply reduced after providers’ results were confidentially disclosed at an institutional level but that public reporting had no added effect on performance (100). Thus performance improvements are accrued through appropriate audit, but in certain healthcare environments there may be little further gain from the extra step of public reporting. One way to try to tackle the possibility of inadvertent or deliberate errors in the recording of risk factors is to publicly report risk factor prevalence alongside the usual risk-​adjusted outcome information. Plots of individual risk factor prevalence are part of the public reports of both PCI and CABG outcomes in the UK.

Strategies to mitigate problems with public reporting Increasing transparency and the associated features of public reporting have many important benefits. Recognizing the potential pitfalls is the key in trying to optimize the benefits, while mitigating potential damage to patient care.

Exclusion of subgroups from public reports An important method is to avoid including certain patient groups in the analyses of publicly reported outcomes. In New York State, there was a 30% fall in the number of patients with cardiogenic shock being treated by PCI between 1997 and 2003. As a result, these patients were finally excluded from analysis for public reporting in 2008, and the use of PCI in this high-​risk population has subsequently increased back to pre-​reporting era levels.

45

46

46

Section 1  

background and basics

The outcomes for patients who experience an ‘out-​of-​hospital cardiac arrest’ are dependent on many factors that are not captured in most PCI datasets. These include time to first cardiopulmonary resuscitation (CPR), presenting rhythm, time to defibrillation for shockable rhythms, bystander CPR, and the time to return of spontaneous circulation. The majority of those who do not survive to hospital discharge die of neurological causes or multiorgan failure from reperfusion injury, and not cardiovascular complications of PCI. The American Heart Association produced a Scientific Statement in 2012 that recommended patients with out-​of-​hospital cardiac arrest be categorized separately and outcomes are tracked, but not publicly reported until adequate risk adjustment models are available (101). In the UK this recommendation has been followed, and such patients are excluded from the public risk-​adjusted outcome reports. In Massachusetts a new ‘exceptional risk’ category was created in 2009 that will capture some of these cases. Patients at ‘exceptional risk’ were defined as being at exceptionally high risk of death but whose risk factors were not collected by their dataset and for whom PCI was their best or only option for improving the chance of survival. All these PCI cases are adjudicated by a committee (www.massdac.org).

Resource allocation towards data collection and validation is mandatory for every strategy

Risk model modifications

References

There are other situations in which conventional datasets and risk models may fail to adjust correctly for risk. In 2005 the Massachusetts Data Analysis Centre added three new covariates to the PCI dataset. These were termed ‘compassionate use’ criteria. They included coma on presentation, use of ventricular assist device, or the use of CPR at the start of the procedure. The allocation of patients to this category is adjudicated by an independent committee, with an appeals process. The impact of adding these novel elements was later analysed and an improvement in the discrimination of the risk model demonstrated (102). It is important that both operators and patients are confident that any model used to adjust for risk is appropriately calibrated. Models need to be continually reviewed to check calibration and discrimination, particularly at the higher end of the spectrum of patient risk.

Disease-​specific outcomes The current focus on procedure-​based analyses can provide a distorted impression of healthcare delivery. This problem was highlighted by the recent analysis of outcomes after presenting with acute myocardial infarction (96). Risk avoidance may result in lower risk patients selected for treatment by PCI. However, the key metric should be the outcome of all patients presenting with myocardial infarction—​not just those treated by PCI. The study from North America implies that, by avoiding treating the sicker patients by PCI, overall mortality was increased. A move towards disease-​specific outcomes might remove some of the drivers of risk-​averse behaviour, the balance being restored by the knowledge that the outcome of patients rejected for an intervention will also be used in assessing the overall quality of a service provided by that individual or that institution. The selection of measures of healthcare performance is important. A narrow focus risks distorting care in other (unmeasured) areas. Procedures may be completed to minimize acute measured complications but without optimizing long-​term consequences. While early mortality is clearly an important metric, it will be important to broaden the scope in future years, to include not only longer term outcomes, but also measures of quality of life—​particularly important in patients with stable coronary artery disease for whom symptom relief is still the dominant reason for revascularization.

Conclusions Selecting patients for procedures is a complex process. As the research base grows, indications for different interventions in a variety of clinical settings are better defined. Interventional techniques evolve and results of earlier trials become redundant. Against this dynamic background risk assessment must provide patients and physicians with the tools to select optimal management strategies and continually improve delivery of care. Providers may ‘game’ the models and patients will arrive with risk factors that are not accounted for adequately by any method of risk stratification. Increasing transparency with publication of unit-​and operator-​ specific outcomes will become more widespread and sophisticated. Despite the complexities of measuring how well we are treating our patients, such measurement is the only way we can drive improvement. We must navigate a course that never compromises our clinical responsibilities to individual patients.

1. Kennedy I. Learning from Bristol: report of the public inquiry into children’s heart surgery at Bristol Royal Infirmary 1984–​1995. BRI Inquiry Panel. London: The Stationery Office, 2001. doi:10.1080/​02688690220148815 2. Berwick D. A Promise to Learn—​A Commitment to Act: Improving the Safety of Patients in England. London: Department of Health, 2013. https://​www.gov.uk/​government/​publications/​ berwick-​review-​into-​patient-​safety 3. Department of Health. The NHS Plan: A Plan for Investment, a Plan for Reform. London: Department of Health, Crown Copyright, 2000. doi:10.1136/​bmj.321.7257.315 4. Darzi A. High Quality Care For All. London: The Stationery Office, 2008. https://​www.gov.uk/​government/​uploads/​system/​uploads/​ attachment_​data/​file/​228836/​7432.pdf 5. Smith J. Fifth Report—​Safeguarding Patients: Lessons from the Past—​ Proposals for the Future. The Shipman Inquiry, 2004. http://​webarchive. nationalarchives.gov.uk/​20090808155110/​http://​www.the-​shipman-​ inquiry.org.uk/​reports.asp 6. Chief Medical Officer. Good Doctors, Safer Patients. London: Department of Health, 2006. 7. Shewhart WA. Economic Control of Quality Of Manufactured Product. Van Nostrand Reinhold, 1931. 8. Shewhart WA. The application of statistics as an aid in maintaining quality of a manufactured product. J Am Stat Assoc. 1925;20:546–​8. 9. Tonino PA, et al. Fractional flow reserve versus angiography for guiding percutaneous coronary intervention. N Engl J Med 2009;360:213–​24. 10. De Bruyne B, et al. Fractional flow reserve–​guided PCI for stable coronary artery disease. N Engl J Med 2014;371:1208–​17. 11. Hannan EL, Kilburn H, Lindsey ML, Lewis R. Clinical versus administrative data bases for CABG surgery. Does it matter? Med Care 1992;30:892–​907. 12. Ugolini C, Nobilio L. Risk adjustment for coronary artery bypass graft surgery: an administrative approach versus EuroSCORE. Int J Qual Health Care 2004;16:157–​64. 13. Lemeshow S, Hosmer Jr, DW. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982;115:92–​106. 14. Marcin JP, Romano PS. Size matters to a model’s fit. Crit Care Med 2007;35:2212–​13. 15. Hanley JA, Meengs WL. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Diagn Radiol 2009;143:29–​36.

47

Chapter 3 

16. Stephan C, Wesseling S, Schink T, Jung K. Comparison of eight computer programs for receiver-​operating characteristic analysis. Clin Chem 2003;49:433–​9. 17. Nashef SAM, et al. European system for cardiac operative risk evaluation (Euro SCORE). Eur J Cardiothorac Surg 1999;16:9–​13. 18. Roques F, et al. Does EuroSCORE work in individual European countries? Eur J Cardio-​Thoracic Surg 2000;18:27–​30. 19. Pitkanen O, Niskanen M, Rehnberg S, Hippelainen M, Hynynen M. Intra-​institutional prediction of outcome after cardiac surgery: comparison between a locally derived model and the EuroSCORE. Eur J Cardio-​Thoracic Surg 2000;18:703–​10. 20. Kawachi Y, et al. Evaluation of the quality of cardiovascular surgery care using risk stratification analysis according to the EuroSCORE additive model. Circ J 2002;66:145–​8. 21. Nashef SA, et al. Validation of European System for Cardiac Operative Risk Evaluation (EuroSCORE) in North American cardiac surgery. Eur J Cardio-​Thoracic Surg 2002;22:101–​5. 22. Jin R, Grunkemeier GL. Additive vs. logistic risk models for cardiac surgery mortality. Eur J Cardiothorac Surg 2005;28:240–​3. 23. Bridgewater B, et al. Surgeon specific mortality in adult cardiac surgery: comparison between crude and risk stratified data. BMJ 2003;327:13–​17. 24. Dewey TM, et al. Reliability of risk algorithms in predicting early and late operative outcomes in high-​risk patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg 2008;135:180–​7. 25. Bhatti F, et al. The logistic EuroSCORE in cardiac surgery: how well does it predict operative risk? Heart 2006;92:1817–​20. 26. Nashef SAM, et al. EuroSCORE II. Eur J Cardio-​Thoracic Surg 2012;41:1–​12. 27. Grant SW, et al. How does EuroSCORE II perform in UK cardiac surgery; an analysis of 23 740 patients from the Society for Cardiothoracic Surgery in Great Britain and Ireland National Database. Heart 2012;98:1568–​72. 28. Barili F, et al. Does EuroSCORE II perform better than its original versions? A multicentre validation study. Eur Heart J 2013;34:22–​9. 29. Howell NJ, et al. The new EuroSCORE II does not improve prediction of mortality in high-​risk patients undergoing cardiac surgery: a collaborative analysis of two European centres. Eur J Cardiothorac Surg 2013;44:1006–​11. 30. Xue Q-​L. NIH Public Access. Clin Geriatr Med 2011;27:1–​15. 31. Afilalo J, et al. Addition of frailty and disability to cardiac surgery risk scores identifies elderly patients at high risk of mortality or major morbidity. Circ Cardiovasc Qual Outcomes 2012;5:222–​8. 32. Ludman PF, et al. Transcatheter aortic valve implantation in the UK: temporal trends, predictors of outcome and 6 year follow up: a report from the UK TAVI Registry 2007 to 2012. Circulation 2015;131:1181–​90. 33. Capodanno D, et al. A simple risk tool (the OBSERVANT score) for prediction of 30-​day mortality after transcatheter aortic valve replacement. Am J Cardiol 2014;113:1851–​8. 34. Iung B, et al. Predictive factors of early mortality after transcatheter aortic valve implantation: individual risk assessment using a simple score. Heart 2014;100:1016–​23. 35. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the aged. The index of ADL. A standardised measure of biological and psychosocial function. JAMA 1963;185:914–​19. 36. Puls M, et al. Impact of frailty on short-​and long-​term morbidity and mortality after transcatheter aortic valve implantation: risk assessment by Katz Index of activities of daily living. EuroIntervention 2014;10:609–​19. 37. Rockwood K, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ 2005;173:489–​95. 38. Seiffert M, et al. Development of a risk score for outcome after transcatheter aortic valve implantation. Clin Res Cardiol 2014;103:631–​40. 39. Rockwood K, et al. A brief clinical instrument to classify frailty in elderly people. Lancet 1999;353:205–​6.

risk assessment and analysis of outcomes

40. Moscucci M, et al. Simple bedside additive tool for prediction of in-​hospital mortality after percutaneous coronary interventions. Circulation 2001;104:263–​8. 41. Shaw RE, et al. Development of a risk adjustment mortality model using the American College of Cardiology—​National Cardiovascular Data Registry (ACC—​NCDR) experience: 1998–​2000. JACC 2002;39:1104–​12. 42. Grayson AD, et al. Multivariate prediction of major adverse cardiac events after 9914 percutaneous coronary interventions in the north west of England. Heart 2006;92:658–​63. 43. Wu C, et al. A risk score to predict in-​hospital mortality for percutaneous coronary interventions. J Am Coll Cardiol 2006;47:654–​60. 44. Valgimigli M, et al. Cyphering the complexity of coronary artery disease using the syntax score to predict clinical outcome in patients with three-​vessel lumen obstruction undergoing percutaneous coronary intervention. Am J Cardiol 2007;99:1072–​81. 45. Chowdhary S, Ivanov J, Mackie K, Seidelin PH, Dzavik V. The Toronto score for in-​hospital mortality after percutaneous coronary interventions. Am Heart J 2009;157:156–​63. 46. Hannan EL, et al. The New York State risk score for predicting in-​hospital/​30-​day mortality following percutaneous coronary intervention. JACC Cardiovasc Interv 2013;6:614–​22. 47. Singh M, et al. Correlates of procedural complications and a simple integer risk score for percutaneous coronary intervention. J Am Coll Cardiol 2002;40:387–​93. 48. De Luca G, et al. Prognostic assessment of patients with acute myocardial infarction treated with primary angioplasty: implications for early discharge. Circulation 2004;109:2737–​43. 49. Addala S, et al. Predicting mortality in patients with ST-​elevation myocardial infarction treated with primary percutaneous coronary intervention (PAMI risk score). Am J Cardiol 2004;93:629–​32. 50. Kunadian B, et al. External validation of established risk adjustment models for procedural complications after percutaneous coronary intervention external validation of established risk adjustment models for procedural complications after percutaneous coronary intervention. Heart 2008;94:1012–​18. 51. Brennan JM, et al. Enhanced mortality risk prediction with a focus on high-​risk percutaneous coronary intervention: results from 1,208,137 procedures in the NCDR (National Cardiovascular Data Registry). JACC Cardiovasc Interv 2013;6: 790–​9. 52. Romagnoli E, et al. EuroSCORE as predictor of in-​hospital mortality after percutaneous coronary intervention. Heart 2009;95:43–​8. 53. Singh M, et al. Mayo Clinic Risk Score for percutaneous coronary intervention predicts in-​hospital mortality in patients undergoing coronary artery bypass graft surgery. Circulation 2008;117:356–​62. 54. Ellis SG, Guetta V, Miller D, Whitlow PL, Topol EJ. Relation between lesion characteristics and risk with percutaneous intervention in the stent and glycoprotein IIb/​IIIa era: an analysis of results from 10,907 lesions and proposal for new classification scheme. Circulation 1999;100:1971–​6. 55. Krone RJ, et al. A simplified lesion classification for predicting success and complications of coronary angioplasty. Registry Committee of the Society for Cardiac Angiography and Intervention. Am J Cardiol 2000;85:1179–​84. 56. Sianos G, et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention 2005;1:219–​27. 57. Serruys PW, et al. Arterial Revascularisation Therapies Study Part II—​ sirolimus-​eluting stents for the treatment of patients with multivessel de novo coronary artery lesions. EuroIntervention 2005;1:147–​56. 58. Leaman DM, Brower RW, Meester GT, Serruys P, van den BM. Coronary artery atherosclerosis: severity of the disease, severity of angina pectoris and compromised left ventricular function. Circulation 1981;63:285–​99. 59. Lansky AJ, Popma JJ. Qualitative and quantitative angiography. In EJ Topol (ed.) Textbook of Interventional Cardiology. Philadelphia, Pennsylvania: Saunders, 1999, pp. 725–​47.

47

48

48

Section 1  

background and basics

60. Lefevre T, et al. Stenting of bifurcation lesions: classification, treatments, and results. Catheter Cardiovasc Interv 2000;49:274–​83. 61. Mohr FW, et al. Coronary artery bypass graft surgery versus percutaneous coronary intervention in patients with three-​vessel disease and left main coronary disease: 5-​year follow-​up of the randomised, clinical SYNTAX trial. Lancet 2013;381:629–​38. 62. Task A, et al. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-​ segment elevation. Eur Heart J 2016;37:267–​315. 63. Amsterdam EA, et al. 2014 AHA/​ACC Guideline for the management of patients with non-​st-​elevation acute coronary syndromes: a report of the American College of Cardiology/​American Heart Association Task Force on Practice Guidelines. Circulation 2014;130:e344–​e426. 64. Farooq V, et al. Anatomical and clinical characteristics to guide decision making between coronary artery bypass surgery and percutaneous coronary intervention for individual patients: development and validation of SYNTAX score II. Lancet 2013;381:639–​50. 65. Chieffo A, et al. Drug-​eluting stent for left main coronary artery disease. The DELTA registry: a multicenter registry evaluating percutaneous coronary intervention versus coronary artery bypass grafting for left main treatment. JACC. Cardiovasc Interv 2012;5:718–​27. 66. Adab P, Rouse AM, Mohammed MA, Marshall T. Performance league tables: the NHS deserves better. BMJ 2002;324:95–​8. 67. Sterne JAC, Smith GD, Cox DR. Sifting the evidence—​what’s wrong with significance tests? Another comment on the role of statistical methods. BMJ 2005;322:226–​31. 68. Spiegelhalter DJ. Funnel plots for comparing institutional performance. Stat Med 2005;24:1185–​202. 69. Shine KI, Isom OW. Adult Cardiac Surgery in New York State 1997–​9. New York State Department of Health. 70. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 1995;57:289–​300. 71. Jones HE, Ohlssen DI, Spiegelhalter DJ. Use of the false discovery rate when comparing multiple health care providers. J Clin Epidemiol 2008;61:232–​40. 72. Shine KI. Percutaneous Coronary Interventions (PCI) In New York State, 2002–​2004. New York State Department of Health, 2006. 73. Riddick Jr, JH, Giddings NW. Computerized preparation of average CUSUM charts for clinical chemistry. Clin Biochem 1971;4:156–​61. 74. de Leval MR, Francois K, Bull C, Brawn W, Spiegelhalter D. Analysis of a cluster of surgical failures. Application to a series of neonatal arterial switch operations. J Thorac Cardiovasc Surg 1994;107:914–​23. 75. Rogers CA, et al. Control chart methods for monitoring cardiac surgical performance and their interpretation. J Thorac Cardiovasc Surg 2004;128:811–​19. 76. Lovegrove J, Valencia O, Treasure T, Sherlaw-​Johnson C, Gallivan S. Monitoring the results of cardiac surgery by variable life-​adjusted display. Lancet 1997;350:1128–​30. 77. Poloniecki J, Valencia O, Littlejohns P. Cumulative risk adjusted mortality chart for detecting changes in death rate: observational study of heart surgery. BMJ 1998;316:1697–​700. 78. Spiegelhalter D, Grigg O, Kinsman R, Treasure T. Risk-​adjusted sequential probability ratio tests: applications to Bristol, Shipman and adult cardiac surgery. Int J Qual Health Care 2003;15:7–​13. 79. Barnard GA. Sequential tests in industrial statistics. J R Stat Soc 1946;8 (Suppl): 1–​26. 80. Armitage P. Sequential tests in prophylactic and therapeutic trials. Q J Med 1954;23:255–​74. 81. Wald A. Sequential tests of statistical hypotheses. Ann Math Stat 1946;6:117–​86. 82. Steiner SH, Cook RJ, Farewell VT, Treasure T. Monitoring surgical performance using risk-​adjusted cumulative sum charts. Biostatistics 2000;1:441–​52.

83. Kunadian B, et al. Cumulative funnel plots for the early detection of interoperator variation: retrospective database analysis of observed versus predicted results of percutaneous coronary intervention. BMJ 2008;336:931–​4. 84. Hall J, Roberts T, Belder M. Door-​to-​balloon time in acute myocardial infarction. N Engl J Med 2007;356:1477–​8. 85. Steiner SH, Cook RJ, Farewell VT. Risk-​adjusted monitoring of binary surgical outcomes. Med Decis Making 2001;21:163–​9. 86. Schneider EC, Epstein AM. Influence of cardiac-​surgery performance reports on referral practices and access to care. A survey of cardiovascular specialists [see comment]. N Engl J Med 1996;335:251–​6. 87. Omoigui NA, et al. Outmigration for coronary bypass surgery in an era of public dissemination of clinical outcomes.[see comment]. Circulation 1996;93:27–​33. 88. Werner RM, et al. Racial profiling: the unintended consequences of coronary artery bypass graft report cards. Circulation 2005;111:1257–​63. 89. Waterworth PD, et al. Factors which influence the cardiac surgeon’s decision not to operate on patients referred for consideration of surgery. J Cardiothorac Surg 2008;3:9. 90. Bridgewater B, et al. Has the publication of cardiac surgery outcome data been associated with changes in practice in northwest England: an analysis of 25,730 patients undergoing CABG surgery under 30 surgeons over eight years. Heart 2007;93:744–​8. 91. Moscucci M, et al. Public reporting and case selection for percutaneous coronary interventions: an analysis from two large multicenter percutaneous coronary intervention databases [see comment]. J Am Coll Cardiol 2005;45:1759–​65. 92. Normand ST. Percutaneous Coronary Intervention in the Commonwealth of Massachusetts. Department of Health Care Policy, Harvard Medical School, 2005. www.massdac.org 93. Normand ST. Percutaneous Coronary Intervention in the Commonwealth of Massachusetts. Department of Health Care Policy, Harvard Medical School, 2003. 94. McCabe JM, Joynt KE, Welt FGP, Resnic FS. Impact of public reporting and outlier status identification on percutaneous coronary intervention case selection in Massachusetts. JACC. Cardiovasc Interv 2013;6:625–​30. 95. Apolito RA, et al. Impact of the New York State Cardiac Surgery and Percutaneous Coronary Intervention Reporting System on the management of patients with acute myocardial infarction complicated by cardiogenic shock. Am Heart J 2008;155:267–​73. 96. Waldo SW et al. Association between public reporting of outcomes with procedural management and mortality for patients with acute myocardial infarction. J Am Coll Cardiol 2015;65:1119–​26. 97. Resnic FS, Welt FGP. The public health hazards of risk avoidance associated with public reporting of risk-​adjusted outcomes in coronary intervention. J Am Coll Cardiol 2009;53:825–​30. 98. Aggarwal B, et al. Cause of death within 30 days of percutaneous coronary intervention in an era of mandatory outcome reporting. J Am Coll Cardiol 2013;62:409–​15. 99. Green J, Wintfeld N. Report cards on cardiac surgeons. Assessing New York State’s approach. N Engl J Med 1995;332:1229–​32. 100. Guru V, et al. Public versus private institutional performance reporting: what is mandatory for quality improvement? Am Heart J 2006;152:573–​8. 101. Peberdy MA, et al. Impact of percutaneous coronary intervention performance reporting on cardiac resuscitation centers: a scientific statement from the American Heart Association. Circulation 2013;128:762–​73. 102. Resnic FS, et al. Improvement in mortality risk prediction after percutaneous coronary intervention through the addition of a ‘compassionate use’ variable to the National Cardiovascular Data Registry Cath PCI dataset: a study from the Massachusetts Angioplasty Registry. J Am Coll Cardiol 2011;57:904–​11.

49

CHAPTER 4

Vascular access: femoral versus radial Andrew Wiper and David H. Roberts

Introduction Gruentzig performed the first intracoronary balloon angioplasty in 1977. In 1993 intracoronary stent deployment became widely accepted. This was in the era before oral dual antiplatelet therapy. The use of intensive antithrombotic therapy at that time led to a dramatic rise in the incidence of femoral arterial access complications. Campeau had already performed the first transradial coronary angiogram in 1989. As the radial artery runs a more superficial course, haemostasis was easier to achieve in a fully anticoagulated patient. Kiemeneij described the first transradial stent deployment in 1995. In 2013 in the UK, 244,229 diagnostic coronary angiograms and 92,589 percutaneous coronary intervention (PCI) procedures were performed (1). Radial access is growing exponentially worldwide and accounted for around 70% of procedures performed in the UK in 2013 (Fig 4.1).

Femoral access Femoral arterial punctures can be placed into four groups (2) based on their location of vascular entry (Fig 4.2). Group 1) Low puncture—​at or below the bifurcation of the common femoral artery (CFA) into the superficial femoral artery (SFA) and profunda femoral artery (PFA). Group 2) Middle puncture—​ above the femoral bifurcation and below the most inferior border of the inferior epigastric artery (IEA). Group 3) High middle puncture—​at or above the inferior border of the IEA and below the origin of the IEA. Group 4) High puncture—​above the origin of the IEA. The ideal puncture site for the CFA is below the inguinal ligament but above the bifurcation into the PFA and SFA. Using the landmarks of the pubic symphysis and the anterior superior iliac crest, the puncture site is approximately 2 cm below the midpoint. A  radio-​opaque marker such as a pair of scissors or forceps can be used to identify the medial femoral head, which is a very reliable marker for the CFA. The inguinal skin lies below the bifurcation in 70% of patients (particularly obese patients) and should not be used as a landmark for common femoral arterial puncture (Fig 4.3).

Ideally, following femoral sheath insertion but prior to PCI, a femoral angiogram should be performed to assess the site of arterial puncture (although this is rarely done). If above the inferior border of the IEA, then there is an increased risk of retroperitoneal haemorrhage following PCI (particularly with the use of glycoprotein IIb/​IIIa receptor antagonists) and the procedure should be performed via an alternative access site (2).

Radial access The radial arterial access site is prepared aseptically in the usual fashion and a 1-​m board is placed on the catheter table to support the abducted arm. The wrist is sometimes hyperextended over a roll of gauze. The position of the radial artery at the wrist displays significant interindividual variability. The best position for arterial puncture is at least 1 cm proximal to the styloid process at the point of maximal pulsation. A common mistake is to attempt to puncture the radial artery in a more distal position over the flexor skin creases. Cannulation at this point is difficult owing to a more mobile arterial axis and hinders applying a haemostatic device effectively at the end of the procedure. Following intra-​arterial sheath placement, a retrograde radial arteriogram may be performed. A mixture of 5 ml of contrast diluted with 5 ml of blood (to minimize any discomfort with contrast injection) is injected via the radial sheath side port. This aids in identification of any arterial anomalies, thus guiding the operator on potential procedural strategies or indeed ascertaining whether an alternative access site would be more suitable. Performing radial coronary angiography or PCI is often challenging at first, and has a steep learning curve (3, 4). When performed by operators experienced in the femoral approach (>200 cases), after approximately 20 radial procedures (4) the procedure duration, fluoroscopy time, and procedural success rate all improve considerably, although 200 procedures are often needed to become completely familiar and competent in the transradial approach.

Left versus right radial arterial access Catheter advancement via the left radial approach follows a 180–​ 200° curve between the left subclavian artery and the left coronary ostium, in contrast to the 90° angle when approaching from the right subclavian artery when utilizing the right transradial approach. The manipulation of Judkins catheters via the left transradial approach is therefore easier and essentially the same

50

50

Section 1  

background and basics % Cases using Radial Access 2013 UK (Source BCIS) 80 71.2 65.2 58.6

60

51.6 43.

40

34.7 26.9 21.3

20

10.2

15.7

0 2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

Figure 4.1  Percentage of diagnostic coronary angiograms and percutaneous coronary intervention procedures using radial access in the UK. Data sourced from the public reports prepared by National Institute of Cardiovascular Outcomes Research (NICOR) on behalf of BCIS.

as the transfemoral approach. Saphenous vein graft cannulation is often easier via the left versus right transradial approach. The left internal mammery artery (LIMA) graft is clearly easier to engage via the left transradial approach although the distal tip of most

Superficial epigastric Superficial circumflex iliac Common femoral

Superficial external pudic Deep external pudic Internal circumflex

External circumflex Descending ramus of external circumflex

mammary artery catheters is steeply curved, and specifically designed radial catheters are available for LIMA cannulation via the left radial approach. It is possible to engage the LIMA graft via the right transradial approach, although traversing across the aortic arch is technically difficult and often unsuccessful. For a left radial approach the operator is on the left side of the patient during the puncture. Following arterial sheath placement, the operator switches to the right side and the arm placed across the patient’s body, and the procedure thereafter performed in the usual manner.

Limitations and contraindications to femoral access Peripheral vascular disease Approximately 10% of patients with cardiovascular disease will have diffuse atherosclerotic disease (5). In patients with known or

Femoral profunda First perforating Superficial femoral

Femoral head

Second perforating

Third perforating Anastomotica magna

Superior external articular branch of popliteal

Figure 4.2  Arterial tree anatomy.

Superior internal articular branch of popliteal

Skin crease

Figure 4.3  Femoral skin crease and femoral head anatomy in an obese patient.

51

Chapter 4 

presumed peripheral vascular disease (history of intermittent claudication; recent vascular surgical assessment or procedure), the femoral approach has to be used with great caution. Having a very low threshold to use fluoroscopy and hydrophilic-​coated wires to aid guide wire advancement will reduce the risk of arterial trauma.

Lower limb arterial vascular surgery and abdominal aortic aneurysm In patients who have undergone femoral bypass grafting, the vessel can be punctured below the surgical conduit but there is little data on the safety of this technique for arterial access. This approach should therefore be avoided if at all possible. The femoral approach is also relatively contraindicated in patients with abdominal aortic aneurysm (AAA). When traversing the aneurysm, there is a risk of thrombotic disruption or inducing a dissection or even perforation. For these reasons, the femoral approach is best avoided.

Extremes of body mass index Vascular complications are highest in extremely underweight and morbidly obese patients (6). In very obese patients, there is more difficulty both in obtaining arterial access and in compressing the femoral artery against the femoral head post-​procedure to obtain haemostasis. The incidence of femoral vascular complications is, however, lowest in moderately obese patients—​the so-​called ‘obesity paradox’ (6–​8). One possible explanation is that, when an operator is faced with an obese patient, femoral access is obtained more cautiously.

Therapeutic anticoagulation A standard recommendation for patients anticoagulated with warfarin is to discontinue warfarin for 3–​4 days prior to their elective procedure, aiming for an international normalized ratio (INR) level of 3 ml/​kg) (37). In these higher risk groups prevention, in the form of adequate periprocedural hydration, discontinuation of nephrotoxic drugs, and close attention to contrast loads during the intervention, are likely to be the most effective means of avoidance. Use of additional agents such as N-​acetylcysteine and sodium bicarbonate have led to heterogeneous results in numerous studies. However, they appear to do little harm and are commonly used in addition to—​but not instead of—​periprocedural hydration in at-​risk populations (38–​40). The type of contrast agent is also important. Ionic contrast agents have been replaced by low or iso-​osmolar non-​ionic solutions, which are associated with less nephrotoxicity (41). Whether iso-​osmolar solutions have a further advantage over hypo-​osmolar contrast agents in preventing contrast nephropathy remains less certain, although ESC guidelines suggest preference for iso-​osmolar agents in patients with moderate–​severe chronic kidney disease (CKD) (6). A  reduction in the volume of contrast used can also be achieved by careful planning—​and staging—​of multivessel intervention. Ascorbic acid has also been identified as a potential protective agent against CIN in patients with pre-​existing CKD (42). High-​dose statins also have a role in preventing CIN (43) and are recommended in ESC guidelines for patients with moderate–​severe CKD (6).

As the nature of interventional cardiology changes and the average age of patients increases, PCI has become increasingly complex. Technological advances have facilitated the treatment of more difficult lesions and decreased the procedural risk. PCI can be ‘high risk’ because of: a) patient factors (e.g. advanced age, morbid obesity, comorbidity); b) anatomical/​coronary factors (e.g. intervention on the left main stem, bifurcation anatomy, calcified disease); and c) clinical factors (e.g. ST-​elevation MI [STEMI], cardiogenic shock), or a combination of one or more of these features. Patients who have undergone high-​risk PCI with any of these features need close attention in the period following revascularization because early identification and prompt management of complications have a significant impact on survival. The increasing use of primary PCI in the setting of STEMI has led to a change in coronary care, with patients bypassing smaller district hospitals as they are rapidly transferred to larger ‘heart attack centres’. These larger centres should have more sophisticated catheter laboratories and high-​dependency areas capable of multiorgan support with easy access to ventilation, haemofiltration, left ventricular support modalities, and cardiac surgery.

Clinical features and treatment of contrast-​induced nephropathy CIN usually occurs within the first 12–​24 hours following contrast exposure and in the majority of cases will recover within 3–​5 days. Additional insults during the PCI procedure, such as prolonged hypotension or pre-​procedural nephrotoxic drugs, can complicate the picture, with the development of acute tubular necrosis and oliguric renal failure. Most therapy is supportive, with a small proportion requiring permanent renal replacement (350 mmol/​l. The development of late or prolonged renal impairment may suggest renal atheroembolic disease, and may be associated with other evidence of embolic complications. Although rare (65% of average ref LA

69%

6-​mo angiography, 2-​year TLR

IVUS better

TULIP (106)

144

Long lesions >20 mm

Single-​centre randomized

Complete apposition, MLD ≥80% 89% of average ref diameter MSA ≥ distal ref LA

6-​mo angiography, 12-​mo MACE

IVUS better

AVID, Angiography versus IVUS-​Directed stent placement trial; CENIC, Central Nacional de Intervenções Cardiovasculares; CFR, coronary flow reserve; CRUISE, Can Routine Ultrasound Influence Stent Expansion study; DIPOL, Direct Stenting versus Optimal Angioplasty; FFR, fractional flow reserve; IVUS, intravascular ultrasound; LA, lumen area; MACE, major adverse cardiac events; MLD, minimum lumen diameter; mo, month; MSA, minimum stent area; MUSIC, Multicentre Ultrasound guided Stent Implantation in Coronaries; NA, not applicable; OPTICUS, Optimization with ICUS to Reduce Stent Restenosis; PRESTO, Results of Prevention of REStenosis with Tranilast and its Outcomes Trial; ref , reference vessel; RESIST, REStenosis after Intravascular ultrasound Stenting; SIPS, Strategy for IVUS-​Guided PTCA and Stenting; SVG, saphenous vein graft; TLR, target lesion revascularization; TVR, target vessel revascularization; TULIP, Thrombocyte activity evaluation and effects of Ultrasound guidance in Long Intracoronary Stent Placement; VA, vessel area.

in a more recent study by Cheneau and colleagues, suggesting that mechanical factors continue to contribute to stent thrombosis, even in this modern stent era, with optimized antiplatelet regimens (109). Although the use of IVUS in all patients for the sole purpose of reducing thrombosis is clearly not warranted from a cost standpoint, IVUS imaging should be considered in patients at particularly high risk for thrombosis (e.g. slow flow) or in whom the consequences of thrombosis would be severe (e.g. left main coronary artery or equivalent).

MSA, as measured by IVUS, is one of the strongest predictors for both angiographic and clinical restenosis following BMS implantation (110–​113). Kasaoka and colleagues indicated that the predicted risk of restenosis decreases 19% for every 1 mm2 increase in MSA and suggested that stents with MSA >9.0 mm2 have a greatly reduced risk of restenosis (112). In the Can Routine Ultrasound Improve Stent Expansion (CRUISE) trial, IVUS guidance by operator preferences increased MSA from 6.25 to 7.14 mm2, leading to a 44% relative reduction in target vessel revascularization at

157

chapter 11 

the role of intravascular ultrasound in percutaneous coronary intervention

9 months, compared with angiographic guidance alone (102). In the Angiography versus IVUS-​Directed stent placement (AVID) trial, IVUS-​guided stent implantation resulted in larger acute dimensions compared with angiography alone (7.55 versus 6.90  mm2), with no increase in complications, and lower 12-​month target lesion revascularization (TLR) rates, particularly for vessels with high-​grade pre-​stent stenosis (98). However, controversial results were also reported in some IVUS-​guided stent trials (114, 115), presumably owing to differing procedural endpoints for IVUS-​ guided stenting, as well as various adjunctive treatment strategies that were utilized in these trials in response to suboptimal results (Table 11.2). Overall, a meta-​analysis of nine clinical studies (2972 patients) demonstrated that IVUS-​ guided stenting significantly lowers 6-​month angiographic restenosis (odds ratio [OR] 0.75, 95% confidence interval [CI] 0.60–​0.94; P=0.01) and target vessel revascularizations (OR 0.62, 95% CI 0.49–​0.78; P=0.00003), with a neutral effect on death and non-​fatal myocardial infarction (MI) compared to an angiographic optimization (116).

Device-​specific intravascular ultrasound insights In-​stent restenosis is primarily caused by intimal proliferation rather than chronic stent recoil (75, 117). Growth of neointima is generally greatest in the areas with the largest plaque burden (41, 118, 119), and the intimal growth process seems to be more aggressive in patients with diabetes or hyperinsulinaemia (120, 121). In the treatment of in-​stent restenosis, IVUS can be helpful to differentiate pure intimal ingrowth from poor stent expansion. A serial IVUS analysis immediately before and after balloon angioplasty for in-​stent restenosis has demonstrated that, in 1090 consecutive in-​stent restenosis lesions, 38% of lesions had an MSA 50% diameter stenosis) at 8  months had greater reference plaque burden (61% versus 49%, P=0.03) (Fig 11.11A) and a higher overexpansion index (maximum stent area/​reference MLA:  1.8 versus 1.5, P=0.03) at baseline, compared to those without edge stenosis (165). Subsequently, the Stent Deployment Techniques on Clinical Outcomes of Patients Treated with the Cypher Stent (STLLR) trial demonstrated that geographic miss (defined as the Intravascular ultrasound-​guided procedures length of injured or stenotic segment not fully covered by DES) Although MSA is one of the strongest predictors for restenosis fol- had a significant negative impact on both clinical efficacy (target lowing stenting, its diagnostic accuracy is often blunted owing to vessel and lesion revascularization) and safety (MI) at 1 year folvarious amounts of neointimal proliferation seen among patients lowing sirolimus-​eluting stent implantation (166). A  more recent treated with BMS. In the DES era, however, the drugs dramatically study evaluating second-​generation DES has also demonstrated reduce the variability of the biological response (neointimal prolif- that 9-​month angiographic edge restenosis was predicted by post-​ eration) and, therefore, the prognostic value of the MSA is magni- stenting reference segment plaque burden >55% (167). Therefore, fied as a powerful predictor for in-​stent restenosis (153–​158). For complete coverage of reference disease with less aggressive stent instance, in an IVUS study of de novo coronary lesions, sirolimus-​ dilatation is currently recommended. Importantly, however, longer eluting stents showed a stronger correlation between baseline MSA stent length has also been reported to be independently associated and 8-​month MLA, compared to control BMS (155). Similar results with DES restenosis and thrombosis (154, 168). Online IVUS guidance can facilitate both the determination of appropriate stent size have also been demonstrated in the second-​generation DES (158). Currently, most interventional cardiologists rely on a compliance and length as well as optimal procedural endpoint, achieving the chart provided by stent manufacturers for the selection of stent goal of covering significant pathology with reasonable stent expansize and inflation pressure. With this method, however, a recent sion, while anchoring the stent ends in relatively plaque-​free vessel IVUS study has shown that the first-​generation DES (sirolimus segments. A number of studies have provided evidence for the long-​term and paclitaxel) achieved only 75% of predicted minimum stent diameter and 66% of predicted MSA in vivo (136). In addition to benefits of IVUS guidance in DES implantation for both complex the adjunctive use of a high-​pressure, non-​compliant balloon post-​ lesions and unselected patient populations (Fig 11.12; Table 11.3). dilatation (159), the utility of IVUS to assure adequate stent ex- The Revascularization for Unprotected Left Main Coronary Artery pansion cannot be overemphasized in daily practice, particularly Stenosis:  Comparison of Percutaneous Coronary Angioplasty COMPARE) registry when there are clinical risk factors for DES failure (e.g. diabetes, versus Surgical Revascularization (MAIN-​ showed a significantly lower 3-​year mortality in the IVUS-​guided renal failure). guided group (4.7% In this context, plaque composition assessment by pre-​ group as compared with the angiography-​ rank P=0.048) in patients treated with DES interventional IVUS can also provide useful information to assure versus 16.0%, log-​ adequate DES deployment. In particular, identification of calcified (Fig 11.12A) (169). A  multicentre registry of IVUS-​guided DES plaque is important, since the presence, degree, and location of cal- implantation for the treatment of bifurcation lesions, the Korean cium within the target vessel can substantially affect the delivery Bifurcation Registry (COBIS), has reported long-​term clinical outand subsequent deployment of coronary stents (160, 161). One im- comes (death or MI) compared with angiography-​guided stenting portant advantage of online IVUS guidance is the ability to acquire (HR 0.44, Cox model, P=0.04) (Fig 11.12B) (170). For long lesions precise information on calcium deposit within a plaque, such as requiring a stent ≥28  mm, prospective, randomized, open-​label, the extent and distance from the lumen. For example, lesions with multicentre trials demonstrated significantly lower MACE rates at extensive superficial calcium may require rotational atherectomy 1 year in the IVUS-​guided DES implantation compared with angiprior to stenting (160, 162). Conversely, even for the lesion with sig- ography-​guided stent implantation (171, 172). With respect to the nificant calcification on fluoroscopy, IVUS may find the calcifica- optimization strategy for complex coronary lesions, the AVIO trial tion to be distributed in a deep portion of the vessel wall or to have was conducted to establish modern, universal criteria for IVUS a calcium arc 90% of distal reference lumen CSA for small vessel, and no edge dissection

18-​mo MACE

No difference

Kim et al. (227)

758

De novo native, bifurcation (BMS + DES)

Single-​centre registry

Discretion of individual operator practice

—​

4-​year MACE and stent thrombosis

IVUS better

Roy et al. (174)

1768

De novo and restenotic native and SVG

Single-​centre registry

Discretion of individual operator practice

—​

30-​d and12-​mo MACE and stent thrombosis

IVUS better

Youn et al. (228)

341

De novo native, STEMI

Single-​centre registry

Discretion of individual operator practice

—​

3-​year death, MI, TVR, and TLR

No difference

ADAPT-​DES (176)

8583

De novo and restenotic native and SVG

Multicentre registry

Discretion of individual operator practice

—​

2-​year stent thrombosis, MI

IVUS better

AVIO (173)

284

De novo native

Multicentre randomized

SA > ‘AOR’ determined by the size of an optimal post-​ dilation balloon based on media-​to-​media diameter measurements

48%

Post-​procedure MLD, 9-​mo MACE

IVUS better (post-​procedure MLD) No difference (9-​mo MACE)

COBIS (170)

487

De novo native, bifurcation

Multicentre registry

Discretion of individual operator practice

—​

1-​year death or MI

IVUS better

EXCELLENT (229)

1421

De novo native

Multicentre non-​randomized

Discretion of individual operator practice

—​

1-​year MACE, stent thrombosis, and MI

Angio better

IRIS-​DES (230)

3244

De novo native

Multicentre registry

Discretion of individual operator practice

—​

2-​year MACE

IVUS better (in patients with a stent length of >22 mm)

MAIN-​COMPARE 975 (169)

De novo left main (BMS + DES)

Multicentre Registry

Discretion of individual operator practice

—​

3-​year mortality

IVUS better

MATRIX (175)

1504

De novo and restenotic native and SVG

Multicentre registry

Discretion of individual operator practice

—​

2-​year TVF, MACE, stent thrombosis

IVUS better

RESET (171)

543

De novo native, long lesions requiring a stent ³28 mm

Multicentre randomized

Discretion of individual operator practice

—​

1-​year MACE

IVUS better (per-​protocol)

IVUS-​XPL (172)

1323

De novo native, long lesions Multicentre requiring a stent ³28 mm randomized

Discretion of individual operator practice

—​

1-​year MACE

IVUS better

CTO

Discretion of individual operator practice

—​

1-​year cardiac death, MACE

IVUS better

CTO-​IVUS (208) 402

Multicentre randomized

ADAPT-​DES, Assessment of Dual Antiplatelet Therapy With Drug-​Eluting Stents; AOR, achievable optimal result; AVIO, Angiography Versus IVUS Optimization; BMS, bare metal stent; COBIS, Korean Bifurcation Registry; CSA, cross-​sectional area; CTO, chronic total occlusion; CTO-​IVUS, Chronic Total Occlusion InterVention with drUg-​eluting Stents guided by IVUS; d, day; DES, drug-​eluting stent; EXCELLENT, Efficacy of Xience/​Promus versus Cypher in Reducing Late Loss after Stenting; IRIS-​DES, Interventional Cardiology Research In-​cooperation Society–​ Drug-​Eluting Stents; IVUS, intravascular ultrasound; IVUS-​XPL, the Impact of Intravascular Ultrasound Guidance on Outcomes of Xience Prime Stents in Long Lesions; MAIN-​COMPARE, Revascularization for Unprotected Left Main Coronary Artery Stenosis: Comparison of Percutaneous Coronary Angioplasty versus Surgical Revascularization; MACE, major adverse cardiac events; MATRIX, Comprehensive Assessment of Sirolimus-​Eluting Stents in Complex Lesions; MI, myocardial infarction; MLD, minimum lumen diameter; mo, month; MSA, minimum stent area; ref, reference vessel; RESET, Real Safety and Efficacy of a 3-​Month Dual Antiplatelet Therapy Following Zotarolimus-​Eluting Stents Implantation; SA, stent area; STEMI, ST-​segment elevation myocardial infarction; SVG, saphenous vein graft; TLR, target lesion revascularization; TVF, target vessel failure; TVR, target vessel revascularization.

162

162

section 2  

percutaneous coronary intervention-related imaging

Baseline

Specific clinical scenarios

Follow-up Resolved ISA

Baseline ISA (+)

Persistent ISA

Late-acquired ISA remodelling (+)

Baseline ISA (−)

Late-acquired ISA remodelling (−)

Figure 11.13  Classification of incomplete strut apposition (ISA). Baseline ISA can either be resolved (resolved ISA) or remain (persistent ISA) at follow-​up. Late-​ acquired ISA without vessel expansion is typically seen in thrombus-​containing lesions, while late-​acquired ISA with focal, positive vessel remodelling is more characteristic of brachytherapy and drug-​eluting stents.

meta-​analysis also suggested a significantly higher risk of late or very late DES thrombosis in patients with ISA at follow-​up (OR 6.51, P=0.02) (184). The main mechanism of LISA after DES is often focal, positive vessel remodelling, whereas plaque regression or thrombus resolution is the predominant mechanism for LISA after BMS (185). With respect to vessel remodelling, incompletely apposed struts are seen primarily in eccentric plaques, and gaps develop mainly on the disease-​free side of the vessel wall. Thus, the combination of mechanical injury at stent implantation and biological injury by DES components may predispose the vessel wall to chronic, pathological dilatation in the setting of little underlying plaque (186). It remains controversial, however, whether this morphological abnormality independently contributes to the occurrence of stent thrombosis, particularly in modern DES technology. In fact, the incidence of LISA has been significantly reduced in second-​ generation DES, in which a considerable proportion of very late DES failure may be attributed to other biological mechanisms, such as accelerated atherosclerotic changes of in-​stent neointima (the so-​called ‘neoatherosclerosis’) (187). Other IVUS-​detected conditions that may be of importance in DES include non-​uniform stent strut distribution and strut fractures (Fig 11.11B) following implantation (153, 188–​190). Theoretically, both abnormalities can reduce the local drug dose delivered to the arterial wall, as well as mechanical scaffolding of the affected lesion segment. By IVUS, strut fracture is defined as longitudinal strut discontinuity and can be categorised based upon its morphological characteristics: (1) strut separation; (2) strut subluxation; or (3) strut intussusceptions (191). Another proposed classification focuses on potential mechanisms of the strut fracture, categorising them based upon the presence and absence of aneurysm at the fracture site (Type I and II, respectively) (192). Angiographic or IVUS studies have reported the incidence of DES fracture to range between 0.8% and 7.7%, wherein in-stent restenosis or stent thrombosis occurred at 22–​88% (193). The exact incidence and clinical implications of strut fractures remain to be investigated in large clinical studies.

Left main disease In the assessment of left main coronary artery (LMCA) disease, angulations, calcification, or spasm in this location can lead to poor catheter engagement and confounded angiographic interpretation. Several investigators showed that high percentages of patients with angiographically normal LMCA had disease by IVUS (193–​196). Conversely, a recent IVUS study demonstrated that fewer than half of angiographically ambiguous left main stenosis had a significant stenosis (197). This was especially true for ostial LMCA disease, where only 36% of the lesions had a significant stenosis and 41% had plaque burden 100 mg) are associated with significantly higher rates of bleeding (13). Its addition to heparin in early trials of PCI led to a 75% relative risk reduction (RRR) in MACE rate (15). Six months of aspirin maintenance therapy after PCI is supported by the M-​Heart II study, showing less reinfarction compared to placebo (16). The data, including those supporting the benefit of aspirin in vascular disease in general, as well as the benefit after acute MI, and against a background of elevated rates of observed ST in patients in whom aspirin is stopped, all contribute to a strong ‘circumstantial’ case for continuing lifelong aspirin after stents. Gastrointestinal irritation is the most common adverse reaction to aspirin but coadministration of a proton pump inhibitor often allows continued treatment. Hypersensitivity to aspirin is reported in up to 10% of patients undergoing PCI. Aspirin sensitivity can be characterized by respiratory symptoms (e.g. exacerbation of asthma); cutaneous manifestations, including rashes and urticaria; angioedema; or anaphylactoid reactions, although severe reactions are uncommon. Rapid desensitization protocols that take between 3 and 6 h can be instituted, with a success rate of up to 98% (17, 18), thus allowing aspirin to be initiated before PCI in the vast majority of cases, with the exception of primary PCI (PPCI).

Thienopyridines Adenosine diphosphate (ADP) released in large quantities from platelet-​dense granules upon platelet activation is an important secondary mediator crucial to the amplification of platelet responses to a variety of agonists. Clopidogrel, ticlopidine, and prasugrel belong to the thienopyridine family of ADP receptor antagonists, which act via inhibition of the platelet P2Y12 receptor. For ticagrelor this is a reversible binding. In vitro these compounds have no antiplatelet activity per se; however, in vivo they undergo hepatic bioactivation by the cytochrome P450 enzyme system. The resulting active

metabolites form disulphide bridges with the extracellular cysteine residues (Cys17 and Cys270) on the P2Y12 receptor and thereby inactivate the receptor. The presence of a methoxycarbonyl group on clopidogrel and prasugrel molecules affords them greater efficacy as well as a superior safety and tolerability profile compared with ticlopidine. Both ticagrelor and prasugrel act more rapidly than clopidogrel.

Ticlopidine Ticlopidine was the first ADP receptor used in clinical practice and, having been largely superseded, is of historical interest only. It was shown to be effective in the secondary prevention of MI and stroke and in preventing ST when used in combination with aspirin. In the STARS study, MACE, including ST, was significantly lower in patients receiving the aspirin–​ticlopidine combination compared to aspirin and warfarin. Furthermore, DAPT with aspirin–​ticlopidine was consistently superior to aspirin with oral anticoagulation in both medium-​, mixed-​(FANTASTIC and ISAR studies) (19, 20), and high-​risk cohorts (MATTIS) (21) undergoing stent implantation. Furthermore, haemorrhagic complications were less common with DAPT in the ISAR, FANTASTIC, and MATTIS studies and were equivalent to aspirin–​warfarin in STARS (19–​22). Compliance with ticlopidine was greatly limited, though, by its poor tolerability owing to side effects, including rashes, diarrhoea, nausea, vomiting, and abdominal pains. Haematological toxicity was the most important undesirable effect. Neutropenia occurred in 2.4% (23) of patients and, whilst marrow failure related to ticlopidine is mostly reversible, it can be complicated by life-​threatening septicaemia. Rarely, ticlopidine causes aplastic anaemia, which carries a high mortality (24). The use of ticlopidine has now largely become superseded by clopidogrel and other P2Y12 inhibitors owing to their superior efficacy and improved tolerability, but very rarely it may be used when sensitivity to the newer agents makes their use prohibitive.

Clopidogrel The CLASSICS study was the first to demonstrate superior tolerability of clopidogrel compared to ticlopidine. Subsequently, the CAPRIE trial evaluated the long-​term safety and efficacy of clopidogrel against aspirin monotherapy in a large cohort of 19,185 patients with a previous history of CVD (25). This showed a modest reduction in MACE of 8.7% in favour of clopidogrel (P = 0.043), adding further to the view that aspirin alone may not be sufficient for secondary prevention of ischaemic events in particular high-​ risk populations. The landmark CURE study randomized 12,562 patients presenting with NSTEMI within 24 h to aspirin with either a combination of loading with clopidogrel 300 mg followed by maintenance (i.e. 75 mg daily), or placebo (26). The addition of clopidogrel reduced MACE by 20% (absolute risk reduction of 2.1%) at 1  year compared to placebo (P < 0.001), although this was associated with more major but non-​fatal bleeding (26). Analysis of data from a subgroup of patients treated with PCI similarly showed a 31% RRR in cardiovascular death and MI (P = 0.002) in the PCI-​CURE Study (27). Further subanalysis suggested that the excess bleeding was directly related to the dose of aspirin used. Owing to its relatively slow onset of action a loading dose is usually administered in the context of PCI, although there is considerable variability in practice with respect to whether this is

365

Chapter 24 

oral antiplatelet therapies in percutaneous coronary intervention

administered before or at the time of PCI. Data from CREDO and PCI-​CURE support the use of a 300-​mg loading dose at least 8–​ 12 h prior to PCI (27, 28). However, several subsequent studies indicated that a 600-​mg loading dose provides additional benefit. The ARMYDA-​2 trial randomized 255 patients to either 300 mg or 600 mg loading doses of clopidogrel 4–​8 h prior to PCI (29). There was a significant decrease in cardiovascular events in the 600-​mg group with no increase in the haemorrhagic risk. Similarly, a randomized trial of patients with ACS receiving loading doses of clopidogrel at least 12 h prior to PCI also found a significant decrease in cardiovascular events at 1  month in those receiving 600  mg compared to 300  mg (30). These randomized studies were supported by a non-​randomized analysis of 4105 unselected patients undergoing PCI, which confirmed a significant decrease in post-​PCI MACE at 1  month in patients receiving 600  mg, without an increase in bleeding complications (31). These studies have informed international clinical guideline committees who have given clopidogrel a class  I  recommendation for patients undergoing PCI for stable coronary artery disease (SCAD), NSTEMI, and STEMI. In the latter ACS populations, however, based on subsequent studies, the European guidelines recommend that clopidogrel is reserved for those patients where other potent therapies such as prasugrel or ticagrelor are either not available or contraindicated (32). Those who receive clopidogrel in the context of PCI should be appropriately loaded with 600 mg at least 6 h prior to the procedure and continue maintenance therapy at a dose of 75 mg daily for a duration dependent on the clinical indication. The duration of DAPT is, however, the subject of intense debate, as will be discussed in the section on ‘Duration of dual antiplatelet therapy’. Furthermore, there is a large body of evidence that demonstrates a wide variation in individual response to clopidogrel. However, no account of this variability is currently taken in routine clinical practice (see ‘Assessment of response to antiplatelet therapy’ section), mostly since it frequently does not translate into adverse clinical outcomes.

Prasugrel Prasugrel is a third-​ generation thienopyridine which provides more rapid, consistent, and potent platelet inhibition compared to clopidogrel, making it an attractive therapeutic alternative in coronary intervention (33). These properties have been attributed to its rapid absorption and more efficient metabolism (it requires only a single enzyme transformation) to a thiol-​containing active metabolite, a process that is less reliant on the liver cytochrome P450 enzyme system compared to its predecessors (34). In addition to its superior pharmacological profile the JUMBO-​ TIMI 26 study showed no excess of haemorrhagic complications at 30  days compared to clopidogrel in 904 patients undergoing elective or urgent PCI (35). A large phase III trial evaluated the efficacy and safety of this novel thienopyridine in 13,608 moderate-​to high-​risk ACS patients scheduled for PCI (1). The TRITON-​TIMI 38 study demonstrated that prasugrel (60 mg loading dose followed by 10  mg maintenance dose) was associated with lower MACE (9.9% versus 12.1%; P < 0.001), as well as less ST (1.1% versus 2.4%; P < 0.001) at 15 months, compared to clopidogrel (300 mg loading dose and 75mg maintenance dose). However, prasugrel was also associated with higher rates of major bleeding (2.4% versus 1.8%; P = 0.03), including fatal as well as non-​fatal life-​threatening bleeding. Post hoc analysis suggested that prasugrel is associated with no net

clinical benefit and could even cause harm in subgroups of older patients (age ≥75 years), low body weight (550; VN P2Y12 assay: 34%

Clopidogrel 600 mg LD + 150 mg MD for 1 month

MACE at 12 months: 3.1% versus 24.6% (P = 0.01)

Ari et al. 2011 (159) (EFFICIENT)

100% SA

47/​47

VN P2Y12 assay: 46 U

Clopidogrel 600 mg LD + 150 mg MD for 1 month

MACE at 6 months: 0 versus 2.6% (P = 0.03)

Studies with long-​term intervention Wang et al. 2011 (161)

80.1% SA; 19.9% NSTEMI

150/​156

VASP: PRI >50%

Stepwise increase in clopidogrel MD to 375 mg (max.) following testing at 1, 3, 6, 9, and 12 months

MACE at 1 year: 9.3% versus 20.4% (P = 0.008)

Price et al. 2011 (150) (GRAVITAS)

60.2% SA; 39.4% UA/​NSTEMI; 0.4% STEMI

1109/​1105

VN P2Y12 assay: PRU ≥230

Clopidogrel 600 mg LD + 150 mg MD for 6 months

MACE at 6 months: 2.3% versus 2.3% (NS)

Trenk et al. 2012 (151) (TRIGGER-​PCI)

100% SA

212/​211

VN P2Y12 assay: PRU ≥208

Prasugrel 60 mg LD + 10 mg MD for 6 months

MACE at 6 months: 0 versus 0.5% (P value not evaluated)

Collet et al. 2012 (152) (ARCTIC)

73% SA; 27% ACS

1213/​1227

VN Aspirin assay: ARU ≥550; VN P2Y12 assay: PRU ≥235 or 80% receptor occupancy was considered optimal to maximize the benefits of these agents (24). The apparent lack of consistency seen in the results from the large number of clinical trials examining the effects of these different agents possibly reflects the heterogeneity in the levels of inhibition obtained (25). The next section will examine the different commercially available GP IIb/​ IIIa antagonists, their mode of action, and the major clinical trials examining their use.

The glycoprotein IIb/​IIIa receptor inhibitors There are three parenteral GP IIb/​IIIa receptor inhibitors commercially available at present:  abciximab (ReoPro®), eptifibatide (Integrelin®), and tirofiban (Aggrastat®).

Abciximab Complete coronary occlusion Acute myocardial infarction

Incomplete coronary occlusion

Spontaneous lysis, repair, and wall remodelling

Temporary resolution of instability Future high-risk coronary lesion

Unstable angina or non-Q-wave myocardial infarction

Figure 25.1  Vulnerable plaque formation and rupture. Reproduced with permission from Yeghiazarians Y, Braunstein JB, Askari A, et al. Unstable angina pectoris. NEJM 2000;342(2):101–​14. Copyright © 2000 Massachusetts Medical Society. All rights reserved.

focus on the development and use of GPIs in current interventional cardiology.

The glycoprotein IIb/​IIIa receptor (the αIIbβ3 integrin) GP IIb/​IIIa receptors are membrane glycoproteins with the αIIbβ3 integrin unique to the cell surface of platelets and megakaryocytes. Between 80,000 and 100,000 of these receptors are present on each platelet. Following plaque rupture, platelets are activated by coming into contact with the newly exposed subendothelium. Stimuli such as adenosine diphosphate (ADP), 5-​HT, thromboxane A2 (TXA2), thrombin, and collagen activate the platelet through receptors present on its surface (22). Once activated, the surface GPs undergo conformational change that results in a high affinity for binding with fibrinogen, their principal ligand. This fibrinogen then links with GP IIb/​IIIa receptors on other platelets and promotes platelet aggregation and the formation of a platelet plug that precipitates coronary thrombosis. Integrin binding affinity is dynamic and is dependent on the receptor’s conformational status. In the resting state, the affinity for fibrinogen binding is low (23). These changes to the GP IIb/​IIIa receptor represent the final common pathway for platelet aggregation and thrombus formation (Fig 25.2). Agents that block this receptor and prevent the binding of fibrinogen have been developed and extensively researched (Fig 25.3). While there

Abciximab is a large chimeric monoclonal antibody that targets the GP IIb/​IIIa receptor on the platelet surface in either the active or resting state. Although it has a high binding affinity for this receptor, abciximab is non-​specific, also binding the vitronectin receptor on vascular smooth muscle cells and the MAC-​1 receptor on monocytes (26). The drug is administered intravenously at an initial bolus dose of 250 μg/​kg body weight followed by a maintenance dose of 0.125 μg/​kg/​min (maximum 10 μg/​min) for up to 36 h, although it is commonly given as a 12-​h infusion. Following the loading dose, a high proportion of the drug is rapidly taken up by platelet GP receptors with only a small amount remaining free in the plasma (27). It is a large molecule and its strong binding results in a prolonged period of platelet inhibition that may last for up to 48 h after the drug has been discontinued. This makes any reversal of the potent antiplatelet effect of abciximab difficult and can have important clinical implications. There is almost no renal excretion of the drug. It is dissipated slowly by being transferred to unoccupied receptor sites on other platelets, gradually reducing the antiplatelet effect and eventually undergoing protease degradation (28). If rapid reversal of action is required then platelet transfusion is effective in speeding this process. Abciximab is a monoclonal antibody and has the potential for the formation of human antichimeric antibodies (HACAs), which occurs in about 6–​7% of patients (29). Although the significance of this is unclear there is concern that this may predispose to immune-​mediated hypersensitivity reactions and profound thrombocytopaenia on subsequent readministrations of the drug. The prevalence of severe and profound thrombocytopaenia in patients who were readministered abciximab (2.8% and 2%, respectively) was 3–​4 times greater than that observed after first-​time administration (30). Neither of the small-​ molecule GP IIb/​ IIIa inhibitors is immunogenic.

Eptifibatide and tirofiban In contrast to abciximab, eptifibatide and tirofiban are synthetic small-​ molecule GP  IIb/​ IIIa receptor inhibitors that bind with

38

Chapter 25 

A

current status of glycoprotein iib/iiia inhibitors

B Adhesion

C Shape change

Inactive platelet Thrombin

Subendothelium

Endothelial cell

ADP

AMP

NO Prostaglandin

D Secretion, generation of TXA2, procoagulant activity Prothrombinase complex

ADP

CD39

E Loose aggregation

F Stabilization, clot retraction

ADP, PDGF, fibrinogen TXA2

GP Ib-IX

α2β1

α5β1

GP IIb-IIIa

vWf

Collagen

Fibronectin

Fibrinogen

Figure 25.2  A) Circulating platelets are usually kept in an inactive state by prostacyclin and nitric oxide (NO) released by the endothelial cells that line the walls of blood vessels. Endothelial cells also express CD39 on their surface, which inhibits platelet activation by converting adenosine diphosphate (ADP), a potent inducer of platelet activation, into adenosine monophosphate (AMP). B, C) At sites where the blood vessel wall has been injured, the platelets adhere to the exposed subendothelium through interactions between collagen, von Willebrand factor (vWf), and fibronectin and their receptors on the platelets, integrin α2β1, glycoprotein Ib-​IX (GP Ib-​IX), and integrin α5β1, respectively. Both thrombin and ADP cause platelets to change into an active conformation. D) Activated platelets secrete ADP, platelet-​derived growth factor (PDGF), and fibrinogen from storage granules in the platelet, and thromboxane A2 (TXA2), produced by immediate biosynthesis. ADP and TXA2 cause circulating platelets to change shape and become activated. E) Glycoprotein IIb/​IIIa receptors on the surface of activated platelets bind fibrinogen, leading to the formation of fibrinogen bridges between the platelets, resulting in platelet aggregation. This, and the simultaneous formation of a fibrin mesh (not shown), lead to the formation of a platelet thrombus. F) Clot retraction then leads to formation of a stable thrombus. Reproduced with permission from MacMillan Publishers Ltd., Bhatt DL, Topol EJ. Scientific and therapeutic advances in antiplatelet therapy. Nat Rev Drug Discov 2003;2:15–​28 © 2003.

platelets only when they are in the active form. They exhibit differing pharmacokinetics to abciximab by binding the receptor with low affinity but in a much more specific and dose-​dependent manner. This competitive and highly reversible mode of action results in a

Agonist Adenosine Diphosphate, Thrombin, Epinephrine, and others Resting platelet GP IIb/IIIa receptors in ligand-unreceptive state

large proportion of the drug freely circulating in the plasma during steady state. Eptifibatide is a cyclic heptapeptide small molecule and is administered intravenously with an initial bolus dose of 180 μg/​kg body weight followed by a maintenance dose of 2.0 μg/​kg/​

GP IIb/IIIa receptor antagonist

Fibrinogen

Inhibition of platelet aggregation GP IIb/IIIa receptors occupied by antagonists

Aggregating platelets GP IIb/IIIa receptors occupied by fibrinogen, which forms bridges between adjacent platelets

Figure 25.3  Showing mechanism of action of glycoprotein IIb/​IIIa inhibitors. Adapted with permission from Lefkovits J, Plow EF, Topol EJ. Platelet glycoprotein IIb/​IIIa receptors in cardiovascular medicine. N Engl J Med 1995;332:1553–​9. Copyright ©1995 Massachusetts Medical Society. All rights reserved.

383

384

384

Section 5  

adjunctive drug therapies in percutaneous coronary intervention

min for up to 72 h, although it has been used for longer periods of time, up to 96 h. Tirofiban is a non-​peptide tyrosine derivative and is administered intravenously at an initial dose of 0.4 μg/​kg/​min for 30 min followed by a maintenance dose of 0.1 μg/​kg/​min for at least 48 h up to 108 h maximum. Both eptifibatide and tirofiban exhibit first-​order kinetics and are predominantly renally excreted. They have a short action of platelet inhibition compared to abciximab (eptifibatide 2–​4 h; tirofiban 4–​8 h). In the case of the need for rapid reversal, platelet transfusion is of little benefit as a large proportion of the drug is non-​receptor-​bound and the receptor sites on the new platelets are simply bound by the freely circulating drug (Table 25.1).

Clinical use of glycoprotein IIb/​IIIa inhibitors Studying the trials examining GP IIb/​IIIa inhibitor use in PCI is a lesson in the history of interventional cardiology. GPIs have been a constant in an ever-​changing sea and these trials have documented all the major changes in the field over the past 20 years. The original trials in the mid-​1990s had low rates of stenting, almost universal femoral vascular access, no oral thienopyridine loading, and high rates of periprocedural ischaemic and vascular complications. In trials involving patients with ACS, early invasive management for non-​STEMI (NSTEMI) with PCI was actively discouraged, thrombolysis for ST-​elevation was widespread, and concepts such as facilitated PCI and combinations of fibrinolytic drugs and GPIs were in common clinical use. The shift to radial vascular access, the development of second-​generation drug-​eluting stents, the roll-​out of primary PCI for STEMI with short door-​to-​balloon times, mechanical thrombus aspiration, and preloading with powerful P2Y12 inhibitors have completely changed the landscape in which the role of GPIs is now being appraised. Although it is very useful to look back at these early trials, they must be examined in the context of the era in which they were carried out.

Percutaneous coronary intervention in stable coronary artery disease The primary clinical endpoint of the clinical trials examining the role of GP IIb/​IIIa inhibitors in conjunction with PCI has been the impact on the incidence of 30-​day ischaemic composite endpoints. These endpoints generally comprise the clinical events of death, MI, and the need for urgent repeat revascularization procedures. Over 15,000 patients have been studied in several large randomized trials (see Table 25.3). All three agents have been evaluated in this setting, although abciximab has been studied the most thoroughly. There has been considerable variation in the duration of post-​PCI infusions of agents between the different trials, with generally shorter duration used for abciximab (12 h) compared to eptifibatide (18–​ 24 h) or tirofiban (18–​36 h). These differences may explain in part some of the inconsistent results from the trials owing to potentially ‘inadequate’ platelet inhibition during and after PCI.

Abciximab In the Evaluation in Percutaneous Transluminal Coronary Angioplasty to Improve Long-​term Outcome with Abciximab GP IIb/​IIIa Blockade (EPILOG) trial (31), a prospective, double-​blind trial, patients at 69 centres were randomized to receive abciximab with standard-​dose, weight-​adjusted heparin (initial bolus of 100 U per kilogram of body weight); abciximab with low-​dose, weight-​ adjusted heparin (initial bolus of 70 U per kilogram); or placebo with standard-​dose, weight-​adjusted heparin. The primary efficacy endpoint was death from any cause, MI, or urgent revascularization within 30  days of randomization. Interestingly, planned stenting was an exclusion criterion. The trial was terminated at the first interim analysis, with 2792 of the planned 4800 patients enrolled. At 30  days, the composite event rate was 11.7% in the group assigned to placebo with standard-​dose heparin, 5.2% in the group assigned to abciximab with low-​dose heparin (hazard ratio [HR] 0.43; P < 0.001), and 5.4% in the group assigned to abciximab with standard-​dose heparin (HR 0.45; P < 0.001). There were no significant differences among the groups in the risk of major bleeding,

Table 25.1  Summary of glycoprotein IIb/​IIIa inhibitor agents commonly used Abciximab

Eptifibatide

Tirofiban

Class

Monoclonal antibody fragment

Cyclic heptapeptide

Non-​peptide

Mechanism of action

Binds receptor causing steric hindrance and conformational changes

Mimics native protein sequence in receptor

Mimics native protein sequence in receptor

Size

Large (48 kDa)

Small (0.8 kDa)

Small (0.5 kDa)

Affinity and receptor binding

High affinity; non-​specific

Low affinity; highly specific for GP IIb/​ IIIa receptor

Low affinity; highly specific for GP IIb/​IIIa receptor

Duration of platelet inhibition

Long acting 24–​48 h

Short-​acting 2–​4 h

Short acting 4–​8 h

Half-​life

30 min (plasma) 12–​16 h (platelet-​bound)

0.85–​2.8 h (plasma) Seconds (platelet-​bound)

1.2–​2 h (plasma) Seconds (platelet-​bound)

Elimination

Protease degredation

Mostly renal

Renal

Dose

250 μg/​kg bolus + infusion 0.125 μg/​kg/​min (max10 μg/​min) for up to 36 h

180 μg/​kg bolus + infusion 2.0 μg/​kg/​ min for up to 72 h

0.4 μg/​kg/​min for 30 min + maintenance dose of 0.1 μg/​kg/​min for at least 48 h up to 108 h

385

Chapter 25 

although minor bleeding was more frequent among patients receiving abciximab with standard-​dose heparin. The Evaluation of Platelet IIb/​IIIa Inhibitor for Stenting (EPISTENT) trial (19), at 63 hospitals in the USA and Canada, randomly assigned 2399 patients undergoing percutaneous revascularization to stenting plus placebo (n = 809), stenting plus abciximab, a IIb/​IIIa inhibitor (n = 794), or balloon angioplasty plus abciximab (n = 796). Again, the primary endpoint was a combination of death, MI, or need for urgent revascularization in the first 30 days. All patients received heparin, aspirin, and standard pharmacological therapy. Fifty-​seven per cent of patients in the study had unstable angina (UA) or recent MI. The primary endpoint occurred in 10.8% patients in the stent plus placebo group, 5.3% in the stent plus abciximab group (P < 0.001), and 6.9% in the balloon plus abciximab group (P = 0.007). Together, these studies suggest that patients receiving abciximab showed a 4.5–​6.5% absolute reduction in the risk for the 30-​day composite endpoint of death, MI, or urgent revascularization. This translates to a 35–​56% relative risk reduction for the composite endpoint. The majority of the effect was because of the reduction in non-​fatal MI and the need for urgent repeat revascularization. Diabetic patients seemed to derive particular benefit in the EPISTENT trial, where they formed a prespecified subgroup. Within this group the composite endpoint occurred in 25.2% of the stent-​placebo group compared to 13% of the stent-​abciximab group (P < 0.005). A  51% reduction in target vessel revascularization (TVR) at 6 months (P = 0.02) was also noted in the diabetic group receiving abciximab, although these findings have yet to be replicated in other trials. Although there was a trend to a reduction in mortality in EPILOG, only the EPISTENT trial showed a statistically significant reduction in mortality at 1  year follow-​up, with a 60% relative risk reduction in the abciximab plus stent group compared to the other treatment groups (32). Pooled-​data of patients with diabetes mellitus was examined from the EPIC, EPILOG, and EPISTENT trials and abciximab was found to reduce the mortality of diabetic patients to the level of placebo-​treated non-​ diabetic patients (4.5% versus 2.5%). The effect was particularly noted in diabetic patients who were also obese and hypertensive and in those patients with diabetes undergoing multivessel disease (33). However, the ISAR-​SWEET (34) prospective study examined diabetic patients undergoing elective PCI who received a 600-​mg loading dose of clopidogrel and excluded patients with ACS and/​or a visible thrombus. No improvement was found in the abciximab group, possibly because of the beneficial effect of the clopidogrel. The effect of clopidogrel was investigated in the ISAR-​REACT trial (35), which compared abciximab with placebo in a randomized trial in low-​risk patients undergoing PCI who had been loaded with high-​dose clopidogrel (600  mg orally) 2 h before the procedure and then continued the clopidogrel for 3  months (75  mg od). At 30 days, there was no difference between the groups, suggesting that the benefit of additional GP IIb/​IIIa inhibitor therapy in low-​risk PCI patients adequately loaded with clopidogrel may be limited. The study, however, may have been underpowered to demonstrate a benefit of abciximab in such a patient group. In select patients, however, with complex lesions and an unstable periprocedural course abciximab may be useful (36).

Eptifibatide The benefit of eptifibatide to reduce events in PCI was evaluated in the Integrelin to Manage Platelet Aggregation to prevent Coronary

current status of glycoprotein iib/iiia inhibitors

Thrombosis-​II (IMPACT-​II) trial (37). This was a double-​blind, placebo-​controlled trial at 82 centres in the USA, enrolling 4010 patients undergoing elective, urgent, or emergency coronary intervention. Patients were assigned one of three treatments: placebo (n = 1328), a bolus of 135 μg/​kg eptifibatide followed by an infusion of 0.5 μg/​kg/​min for 20–​24 h (n = 1349), or 135 μg/​kg eptifibatide bolus with a 0.75 μg/​kg/​min infusion (n = 1333). The coronary procedure was started within 10–​60 min of the start of study treatment. The primary endpoint was the 30-​day composite occurrence of death, MI, unplanned surgical or repeat percutaneous revascularization, or coronary stent implantation for abrupt closure (by intention to treat). By 30 days, the composite endpoint had occurred in 151 (11.4%) patients in the placebo group compared with 124 (9.2%) in the 135/​0.5 eptifibatide group (P = 0.063), and 132 (9.9%) in the eptifibatide 135/​0.75 group (P = 0.22). By treatment-​received analysis, the 135/​0.5 regimen produced a significant reduction in the composite endpoint (11.6 versus 9.1%; P = 0.035), but the 135/​0.75 regimen did not produce a significant reduction (11.6 versus 10.0%; P = 0.18). A potential reason for the reduced clinical effect seen in the IMPACT-​II trial may have been problems with the assay used to measure platelet activity that resulted in lower than optimal doses of eptifibatide being used (38). The Enhanced Suppression of the Platelet IIb/​IIIa Receptor with Integrelin Therapy (ESPRIT) trial (39) re-​examined the effect of eptifibatide on patients undergoing PCI but used a fourfold higher dose than in IMPACT-​II. In all, 2064 patients were recruited in a multicentre study. Immediately before PCI, patients were randomly allocated to receive eptifibatide, given as two 180 μg/​kg boluses 10 min apart and a continuous infusion of 2.0 μg/​kg/​min for 18–​24 h, or placebo, in addition to aspirin, heparin, and a thienopyridine. The primary endpoint was again the composite of death, MI, urgent TVR, and thrombotic bail-​out GP IIb/​IIIa inhibitor therapy. The trial was terminated early for efficacy. The primary composite endpoint was reduced from 10.5% to 6.6% with treatment (P = 0.0015). The key secondary endpoint (composite of death, MI, or urgent TVR at 30 days) was also reduced, from 10.5% to 6.8% (P = 0.0034). The benefit of eptifibatide was sustained, still being present at 6 months and 1 year (40, 41). The majority of the benefit, like many of these trials, was from reduction in MI and although there was a trend towards a reduction in mortality this was not statistically significant. The INtegrilin plus STenting to Avoid myocardial Necrosis Trial (INSTANT) trial assessed the role of eptifibatide in patients with diffuse coronary disease undergoing drug-​eluting stenting was negative but was stopped early after poor recruitment (42).

Tirofiban There have been few trials examining the use of tirofiban in stable patients undergoing routine PCI. A  small, single-​ centre study involving 93 patients served as a dose-​ranging and safety study (43). The Do Tirofiban And Reopro Give Similar Efficacy Outcomes Trial (TARGET) (44) used a double-​ blind, double-​ dummy design at 149 hospitals in 18 countries and was designed and statistically powered to demonstrate the non-​inferiority of tirofiban as compared with abciximab. Patients were randomly assigned to receive either tirofiban or abciximab before undergoing PCI with the intent to perform stenting. The primary endpoint was a composite of death, non-​fatal MI, or urgent TVR at 30 days. The primary endpoint occurred more frequently among the 2398 patients in the tirofiban group than among the 2411 patients in

385

386

386

Section 5  

adjunctive drug therapies in percutaneous coronary intervention

the abciximab group (7.6% versus 6.0%; HR 1.26; one-​sided 95% confidence interval [CI] 1.51, demonstrating lack of equivalence, and two-​sided 95% CI 1.01–​1.57, demonstrating the superiority of abciximab over tirofiban; P = 0.038). The majority of the difference in composite endpoint was due to a lower incidence of MI. The relative benefit of abciximab was consistent regardless of age, sex, the presence or absence of diabetes, or the presence or absence of pretreatment with clopidogrel. There were no significant differences in the rates of major bleeding complications or transfusions, but tirofiban was associated with a lower rate of minor bleeding episodes and thrombocytopaenia. It was suggested that an insufficient dosing regime and patients failing to reach adequate levels of platelet inhibition perhaps attenuated the clinical benefit of tirofiban in this study. This issue was due to be addressed in the TENACITY trial (45), a multicentre study comparing tirofiban at a higher dose to abciximab in patients undergoing PCI, but this was stopped early due to lack of funding.

Conclusion The early trials that demonstrated GP IIb/​IIIa inhibitor benefit in stable coronary disease had low rates of stenting and thienopyridine use, and it is therefore difficult to apply these data to a contemporary PCI population. More recent trials with routine preloading with thienopyridine and planned coronary stenting did not demonstrate additional benefit from GP IIb/​IIIa inhibitors and therefore routine use in stable coronary disease is not recommended by the latest guidelines (Box 25.1). Anecdotal experience, however, suggests that GP IIb/​IIIa inhibitors may be beneficial in ‘bail-​out’ situations (intraprocedure thrombus formation, slow flow, threatened vessel closure) (46).

Box 25.1  Showing the current guidelines for the use of glycoprotein IIb/​IIIa inhibitors

Current guidelines for stable patients undergoing percutaneous coronary intervention (PCI)* American College of Cardiology/​American Heart Association/​ Society for Cardiovascular Angiography and Interventions (ACC/​AHA/​SCAI Practice Guidelines (2011) (121) CLASS IIa In patients undergoing elective PCI treated with unfractionated heparin (UFH) and not pretreated with clopidogrel, it is reasonable to administer a glycoprotein (GP) IIb/​ IIIa inhibitor (abciximab, double-​ bolus eptifibatide, or high-​ bolus dose tirofiban) (Level of Evidence: B) * Class I: Evidence and/​or general agreement that a given diagnostic procedure/​treatment is beneficial, useful, and effective; Class II: Conflicting evidence and/​or a divergence of opinion about the usefulness/​efficacy of the treatment; Class IIa: Weight of evidence/​opinion is in favour of usefulness/​ efficacy; Class IIb: Usefulness/​efficacy is less well established by evidence/​ opinion. Level of evidence A: Data derived from multiple randomized clinical trials or meta-​analyses; Level of evidence B: Data derived from a single randomized clinical trial or large non-​randomized studies; Level of evidence C: Consensus of opinion of the experts and/​or small studies, retrospective studies, registries.

CLASS IIb In patients undergoing elective PCI with stent implantation treated with UFH and adequately pre-​treated with clopidogrel, it might be reasonable to administer a GP  IIb/​IIIa inhibitor (abciximab, double-​bolus eptifibatide, or high-​bolus dose tirofiban). (Level of Evidence: B) European Society of Cardiology Guidelines (2017) (46) Class  IIa. GP IIb/​IIIa inhibitors should be considered only for bail-​out. (Level of evidence: C)

Current guidelines for patients with non-​ST elevation acute coronary syndromes ACC/​AHA/​SCAI Practice Guidelines (2014) (122) Class  I. In patients with non-​ST-​elevation acute coronary syndrome (NSTEACS) and high-​risk features (e.g. elevated troponin) who are not adequately pre-​treated with clopidogrel or ticagrelor, it is useful to administer a GP IIb/​IIIa inhibitor (abciximab, double-​bolus eptifibatide, or high-​dose bolus tirofiban) at the time of PCI. (Level of evidence: A) Class IIa. In patients with NSTEACS and high-​risk features (e.g. elevated troponin) treated with UFH and adequately pre-​treated with clopidogrel, it is reasonable to administer a GP IIb/​IIIa inhibitor (abciximab, double-​bolus eptifibatide, or high-​dose bolus tirofiban) at the time of PCI. (Level of evidence: B) ESC Guidelines (2015) (123) Class IIa. GP IIb/​IIIa inhibitors during PCI should be considered for bail-​out situations or thrombotic complications. (Level of evidence: C) Class  III. It is not recommended to administer GP IIb/​IIIa inhibitors in patients in whom coronary anatomy is not known. (Level of evidence: A)

Current guidelines for patients undergoing primary PCI for ST-​elevation myocardial infarction (STEMI) ACC/​AHA/​SCAI Practice Guidelines (2013) (127) Class IIa. It is reasonable to begin treatment with an intravenous GP IIb/​IIIa receptor antagonist such as abciximab (Level of evidence:  A), high-​ bolus-​ dose tirofiban (Level of evidence:  B), or double-​bolus eptifibatide (Level of evidence:  B) at the time of primary PCI (with or without stenting or clopidogrel pre-​ treatment) in selected patients with STEMI who are receiving unfractionated heparin. Class IIb. 1. It may be reasonable to administer intravenous GP IIb/​IIIa receptor antagonist in the pre-​catheterization laboratory setting (e.g. ambulance, emergency department) to patients with STEMI for whom primary PCI is intended. (Level of evidence: B) 2. It may be reasonable to administer intracoronary abciximab to patients with STEMI undergoing primary PCI. (Level of evidence: B) ESC Guidelines (2017) (46) Class IIa. GP IIb/​IIIa inhibitors should be considered for bail-​out or evidence of no-​reflow or a thrombotic complication. (Level of evidence: C).

387

Chapter 25 

Non ST-​elevation acute coronary syndromes A large number of patients have been studied in trials assessing the benefit of administration of GP IIb/​IIIa inhibitors in ACS but again the majority of these trials do not follow current clinical practice, with low rates of stenting and dual oral antiplatelet loading. These studies have utilized many different trial designs and definitions for clinical endpoints that must be considered in their evaluation. In some trials PCI was planned per protocol (CAPTURE [47], EPIC [32], RESTORE [48], ISAR-​REACT 2 [49], PRISM-​PLUS [50]) and in others an invasive strategy was discouraged (GUSTO-​IV [51], PRISM [52], and PARAGON [53]) or left to the discretion of the operator (PURSUIT [54]). Even in the older studies with planned PCI, stent rates were generally low, only 7.6% in CAPTURE and below 2% in EPIC and RESTORE where stenting was discouraged.

Glycoprotein IIb/​IIIa inhibitors without planned percutaneous coronary intervention in acute coronary syndrome In the GUSTO-​IV ACS trial, 7800 ACS patients treated with aspirin and either unfractionated heparin (UFH) or low-​molecular weight heparin (LMWH) were randomized to an abciximab bolus and 24-​h infusion, an abciximab bolus and 48-​h infusion, or placebo. The trial found no reduction in the composite endpoints at 30 days and an increase in bleeding from abciximab treatment (51). The PURSUIT trial was designed to allow investigators to treat patients according to institutional standard practice without mandated revascularization protocols. A  total of 10,948 ACS patients was randomly assigned, in a double-​blind manner, to receive a bolus and infusion of either eptifibatide or placebo, in addition to standard therapy, for up to 72 h (or up to 96 h, if coronary intervention was performed near the end of the 72-​h period). The primary endpoint was a composite of death and non-​fatal MI occurring up to 30 days after the index event. As compared with the placebo group, the eptifibatide group had a 1.5% absolute reduction in the incidence of the primary endpoint (14.2% versus 15.7% in the placebo group; P = 0.04). The benefit was apparent by 96 h and was still apparent at 30 days. Benefits were most marked in those that underwent early PCI within the first 72 h (n = 1228), with a 31% reduction in the incidence of the composite endpoint of death or non-​fatal MI at 30 days in those treated with eptifibatide, as compared to placebo (11.6% versus 16.6%; P = 0.01) (54). Indeed, at 30  days eptifibatide had no significant effect over placebo in patients who did not receive early PCI*. In the Platelet Receptor Inhibition in Ischemic Syndrome Management (PRISM) trial, a double-​blind study, 3232 patients who were already receiving aspirin were randomly assigned to additional treatment with intravenous tirofiban for 48 h. The primary endpoint was a composite of death, MI, or refractory ischaemia at 48 h. The incidence of the composite endpoint was 32% lower at 48 h in the group that received tirofiban (3.8% versus 5.6% with heparin; risk ratio [RR] 0.67, 95% CI 0.48–​0.92; P = 0.01). Percutaneous revascularization was only performed in 1.9% of the patients during the first 48 h. At 30 days, the frequency of the composite endpoint (with the addition of readmission for unstable angina) was similar in the two groups (15.9% in the tirofiban group versus 17.1% in

*Some caution must be expressed in this subgroup analysis, as the patients were selected for early PCI in a non-​randomized fashion.

current status of glycoprotein iib/iiia inhibitors

the heparin group; P = 0.34). There was a trend toward a reduction in the rate of death or MI with tirofiban and mortality was 2.3%, as compared with 3.6% in the heparin group (P = 0.02). Major bleeding occurred in 0.4% of the patients in both groups. The Platelet IIb/​IIIa Antagonism for the Reduction of Acute coronary syndrome events in a Global Organization Network (PARAGON A) study tested the benefit of different doses of lamifiban (an alternative small-​molecule GP IIb/​IIIa inhibitor) alone and in combination with heparin in ACS patients undergoing initial medical therapy without revascularization unless clinically necessitated (53). The composite primary endpoint of death or non-​fatal MI at 30  days was no different between the groups compared to standard therapy. By 6  months, this composite was lowest for those assigned to low-​dose lamifiban (P = 0.027) and intermediate for those assigned to high-​dose lamifiban (P = 0.450) compared with control (13.7%, 16.4%, and 17.9%, respectively). A large meta-​analysis of six trials that included 31,402 patients with ACS for whom an interventional strategy was not planned found a 9% relative reduction in the odds of death or MI at 30 days in patients treated with GP IIb/​IIIa inhibitors compared to placebo (10.8 versus 11.8%; odds ratio [OR] 0.91; 95% CI 0.84–​0.98; P = 0.015) (55). The event reduction was greatest in patients at greatest risk of thrombotic complications, as suggested by raised troponin levels. Major bleeding complications were increased by treatment with GP IIb/​IIIa inhibitors (2.4% versus 1.4%; P < 0.0001).

Glycoprotein IIb/​IIIa inhibitors with planned percutaneous coronary intervention in acute coronary syndrome The Chimeric c7E3 FAB Antiplatelet Therapy in Unstable Refractory Angina (CAPTURE) trial examined 1265 patients with refractory unstable angina to investigate whether serum troponin T levels identify patients most likely to benefit from treatment with abciximab (47). Serum troponin T levels at the time of study entry were elevated (above 0.1 ng per ml) in 275 patients (21.7%). Among patients receiving placebo, the risk of death or non-​fatal MI was related to troponin T levels. The 6-​month cumulative event rate was 23.9% among patients with elevated troponin T levels, as compared with 7.5% among patients without elevated troponin T levels (P < 0.001). Among patients treated with abciximab, the respective 6-​ month event rates were 9.5% for patients with elevated troponin T levels and 9.4% for those without elevated levels. As compared with placebo, the RR of death or non-​fatal MI associated with treatment with abciximab in patients with elevated troponin T levels was 0.32 (95% CI 0.14–​0.62; P = 0.002). The lower event rates in patients receiving abciximab were attributable to a reduction in the rate of MI (OR 0.23; 95% CI 0.12–​0.49; P < 0.001). In patients without elevated troponin T levels, there was no benefit of treatment with respect to the relative risk of death or MI at 6 months (OR 1.26; 95% CI 0.74–​2.31; P = 0.47). This study suggests that serum troponin T level can be used to identify a high-​risk subgroup of patients with ACS who are scheduled for PCI who are likely to benefit particularly from treatment with abciximab (47). The Evaluation of c73 for the Prevention of Ischaemic Complications (EPIC) trial was a prospective, randomized, double-​ blind trial involving 2099 patients treated at 56 centres scheduled to undergo coronary angioplasty or atherectomy in high-​risk clinical situations involving severe UA, evolving acute MI, or high-​risk coronary morphological characteristics (29, 56). Abciximab was

387

38

388

Section 5  

adjunctive drug therapies in percutaneous coronary intervention

compared to placebo and patients received a bolus and an infusion of drug. At 30  days, a 35% relative reduction in the rate of the primary composite endpoint (12.8% versus 8.3%; P = 0.008) was demonstrated. These effects were persistent and maintained at 3-​year follow-​up (57). The majority of the impact on composite endpoint was due to a reduction in non-​fatal MI and need for urgent revascularization. In the ISAR-​REACT 2 trial, this benefit—​ according to the primary endpoint of death, MI, or urgent TVR within 30 days—​was maintained despite clopidogrel pre-​treatment with a loading dose of 600 mg in patients with NSTEMI (13.1% versus 18.3%; RR 0.71; 95% CI 0.54–​0.95; P < 0.02), but not in UA without biomarker protein elevation (4.6% versus 4.6%; RR 0.99; 95% CI 0.56–​1.76; P = 0.98) (49). The effect of tirofiban in 2139 patients undergoing PCI early after ACS (within 72 h) was examined in the Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis (RESTORE) trial (48). The primary composite endpoint was used, including death from any cause, MI, and urgent TVR for recurrent ischaemia at 30 days. Although patients receiving tirofiban experienced fewer events at 48 h and at 7 days (38% and 27% relative reduction, respectively), the reduction in primary composite endpoint was non-​significant at 30 days. When repeat angioplasty or coronary artery bypass surgery procedures were included in the composite only if performed on an urgent or emergency basis, the composite 30-​day event rates were 10.5% for the placebo group and 8.0% for the tirofiban group, a relative reduction of 24% (P = 0.052), suggesting more benefit in a higher risk population. This was similar to findings from the EPIC and IMPACT-​II studies, where a more pronounced benefit was seen in high-​risk patients. In the RESTORE trial there was criticism that the dosing regimen of tirofiban was inadequate, resulting in suboptimal platelet inhibition. The additive value of tirofiban administered with the high-​dose bolus (HDB) regimen was assessed in the ADVANCE trial (58). Two hundred and two patients undergoing high-​ risk PCI, pre-​ treated with thienopyridines, were consecutively randomized to HDB tirofiban (25 μg/​kg/​3 min, and infusion of 0.15 μg/​kg/​min for 24–​48 h) or placebo immediately before the procedure and then followed for a median time of 185 days (range 45–​324 days) for the occurrence of the primary composite endpoint of death, MI, TVR, and bail-​out use of GPIs. The cumulative incidence of the primary endpoint was 35% and 20% in placebo and HDB tirofiban groups, respectively (HR 0.51; 95% CI 0.29–​0.88; P  =  0.01). This difference was mainly due to the reduction of MI and bail-​out use of GP IIb/​IIIa inhibitors, with no significant effect on TVR or death. The safety profile did not differ between tirofiban and placebo. The issue of the timing of administration of GP IIb/​ IIIa agents in ACS patients undergoing PCI was addressed in the Acute Catheterization and Urgent Intervention Triage Strategy (ACUITY-​Timing) trial (59). In all, 9207 patients with moderate-​ and high-​risk ACS undergoing an invasive treatment strategy were randomly assigned to receive either routine upstream (tirofiban or eptifibatide) or deferred selective (eptifibatide or abciximab) GP IIb/​IIIa inhibitor administration just prior to PCI, respectively. The primary outcome was assessment of non-​inferiority of deferred GP IIb/​IIIa inhibitor use compared with upstream administration for the prevention of composite ischaemic events (death, MI, or unplanned revascularization for ischaemia) at 30  days. Endpoints occurred in 7.9% of patients assigned to deferred use compared with 7.1% of patients assigned to upstream administration that,

while not statistically significant, did not meet the criteria for non-​inferiority. However, upstream use of GP IIb/​IIIa inhibitors resulted in increased rates of major bleeding compared to the deferred group (6.1% versus 4.9%, respectively; P < 0.001) as a result of more frequent usage and longer duration of therapy. Overall, the net clinical benefit was probably similar in both groups. The Early Glycoprotein IIb/​IIIa Inhibition in Patients with Non-​ST-​Segment Elevation Acute Coronary Syndrome (EARLY-​ ACS) trial (60) examined the question of optimal timing of GP IIb/​IIIa administration where early eptifibatide was compared with placebo (with provisional eptifibatide in the catheterization laboratory) in 9492 patients with high-​risk ACS scheduled for early PCI (within 12–​96 h of randomization). The primary endpoint of all-​cause mortality, MI, ischaemia-​ driven revascularization, or thrombotic bail-​ out at 96 h, occurred in 9.3% in the early therapy group compared to 10% in the provisional eptifibatide arm (OR 0.92; 95% CI 0.8–​ 0.16; P = 0.23) with or without clopidogrel therapy. The secondary endpoint of all-​cause mortality or MI at 30 days slightly favoured the early GP IIb/​IIIa inhibitor group (11.2% versus 12.3%; CI 0.79–​ 1.01; P = 0.08), but at the cost of more Thrombolysis in Myocardial Infarction (TIMI) major bleeding (2.6% versus 1.8%; P = 0.02) and increased need for blood transfusion (8.6% versus 6.7%; P = 0.001). It has been suggested that GP IIb/​IIIa inhibitors are most beneficial in high-​risk ACS patients undergoing PCI, and a meta-​analysis by Roffi et al. (61) examined the diabetic populations enrolled in six large-​scale platelet GP IIb/​IIIa inhibitor ACS trials: PRISM, PRISM-​ PLUS, PARAGON A, PARAGON B, PURSUIT, and GUSTO IV. Among 6458 diabetic patients, platelet GP IIb/​IIIa inhibition was associated with a significant mortality reduction at 30 days, from 6.2% to 4.6% (OR 0.74; 95% CI 0.59–​0.92; P = 0.007). Conversely, 23,072 non-​diabetic patients had no survival benefit (3.0% versus 3.0%). The interaction between platelet GP IIb/​IIIa inhibition and diabetic status was statistically significant (P = 0.036). The survival benefit was more apparent among the 1279 diabetic patients undergoing PCI during index hospitalization. The use of these agents was associated with a mortality reduction that was already apparent at 30 days, from 4.0% to 1.2% (OR 0.30; 95% CI 0.14–​0.69; P = 0.002). The ACUITY trial compared a regimen of bivalirudin alone against UFH plus GPIs and found a significant benefit of bivalirudin with respect to the primary 30-​day composite endpoint of ischaemic and bleeding complications (10.1% versus 11.7%, respectively; RR 0.86; 95% CI 0.77–​0.97; P < 0.02), driven by a reduction in major bleeding complications (3.0% versus 5.7%, respectively; RR 0.53; 95% CI 0.43–​0.65; P < 0.001) without a significant increase in ischaemic complications (7.8% versus 7.3%, respectively; RR 1.08; 95% CI 0.93–​1.24; P = 0.32) (62). Bail-​out GP IIb/​IIIa inhibitor usage in the bivalirudin arm was 7.4%. This benefit of bivalirudin was found regardless of whether GPIs were administered downstream or upstream and was maintained during 1-​year follow-​up. In the ISAR-​REACT 4 trial ACS patients undergoing PCI did not significantly benefit from abciximab with UFH compared with bivalirudin alone (63). The primary endpoint of death, recurrent MI, urgent TVR, or major bleeding within 30 days occurred in 10.9% of patients in the abciximab group, as opposed to 11.0% in the bivalirudin group (RR 0.99; 95% CI 0.74–​1.32; P =  0.94). However, heparin plus abciximab was associated with significantly more major bleeding than bivalirudin (4.6% versus 2.6%, respectively; RR 1.84; 95% CI 1.10–​3.07; P = 0.02). In TRITON-​TIMI 38, 7414 patients (54.5% of the total study population) received a

389

Chapter 25 

GP IIb/​IIIa inhibitor and, while prasugrel conferred a consistent benefit in terms of reducing the risk of cardiovascular death, MI, or stroke, when compared with clopidogrel, this was irrespective of the use of GPIs (64). Similarly, the risk of TIMI major or minor bleeding was unaffected with respect to whether or not patients were treated with GPIs.

Conclusion In the era before routine early invasive management with PCI and stenting and the preloading of patients with oral dual antiplatelet therapy, ACS trials demonstrated a lower incidence of composite ischaemic events in favour of GP IIb/​IIIa inhibitor treatment in combination with UFH than with UFH alone, primarily through a reduction in MI, although at the expense of increased bleeding. In the current era there is no evidence for additional benefit from routine upstream use of GPIs in non-​ST elevation ACS patients scheduled for invasive management with PCI who have been adequately loaded with thienopyridines. As in stable coronary disease, anecdotal experience, however, suggests that GPIs may be beneficial in ‘bail-​out’ situations such as propagating intraprocedure thrombus formation, slow-​flow, or threatened vessel closure (46).

ST-​elevation myocardial infarction The cornerstone of therapy for acute STEMI involves the rapid restoration of epicardial bloodflow and myocardial tissue perfusion. Primary PCI is established as the treatment of choice, offering more definitive revascularization in a wider range of patients and with fewer bleeding complications (65). In the UK, like many other developed countries, sophisticated networks of regional Heart Attack Centres have now rolled out round the clock primary PCI coverage offering rapid and effective reperfusion with short ‘pain-​to-​balloon’ times and reduced ischaemia time. In many regions around the world, however, prompt primary PCI is still not available and fibrinolysis remains in widespread use. In addition, even when primary PCI is available no/​slow-​reflow with impaired tissue perfusion remains a problem in up to a third of patients following successful stent implantation. This is the field of contemporary interventional cardiology where GP IIb/​IIIa inhibitors as adjunctive agents to improve outcomes for reperfusion therapy for STEMI still have a significant role as front-​line treatment in certain cases.

Glycoprotein IIb/​IIIa inhibitors as adjunctive therapy to fibrinolysis Fibrinolytic agents fail to achieve optimum reperfusion in many patients with STEMI. The TIMI 14 trial (66) explored the hypothesis that the addition of abciximab would facilitate the rate and extent of thrombolysis when combined with a half dose of fibrinolytic. They examined 888 patients undergoing fibrinolysis with alteplase for STEMI, and found that the addition of abciximab produced early, marked increases in TIMI 3 flow at 90 min compared to alteplase alone. This improvement in reperfusion occurred without an increase in the risk of major bleeding. Modest improvements in TIMI 3 flow were seen when abciximab was combined with streptokinase, but there was a marked increased risk of bleeding. The Strategies for Patency Enhancement in the Emergency Department (SPEED) Group (67) found similar improvements in vessel patency using a combination of abciximab and low-​dose reteplase. Again, there were increased rates of minor bleeding rates in the combination group but major bleeding rates were similar.

current status of glycoprotein iib/iiia inhibitors

The GUSTO V AMI trial (68) assessed the clinical benefits of GP IIb/​IIIa inhibitor as an adjunct to fibrinolysis. This was a randomized, open-​label trial involving 16,588 patients in the first 6 h of evolving STEMI to compare the effect of the standard-​dose reteplase alone with half-​dose reteplase plus abciximab. The primary endpoint was 30-​day mortality, and secondary endpoints included various complications of MI. Analysis was by intention to treat. At 30 days, there was no difference in primary endpoint between the two groups (5.9% versus 5.6%; P = 0.43). In the combination group there was a reduction in the composite endpoint of deaths or non-​fatal reinfarction in the combination group compared to reteplase alone (8.8% versus 7.4%; P = 0.0011). There was also less need for urgent revascularization (8.6% versus 5.6%; P = 0.001) and fewer major non-​fatal ischaemic complications of acute MI. On the other hand, these benefits were partly counterbalanced by increased non-​intracranial bleeding complications in the combination group (4.3% versus 1.9%; P < 0.0001). The rates of intracranial haemorrhage and non-​fatal disabling stroke were similar in both groups. ASSENT-​ 3 randomized 6095 patients with STEMI of less than 6 h to one of three regimens:  full-​dose tenecteplase and enoxaparin for a maximum of 7  days, half-​dose tenecteplase with weight-​adjusted low-​dose UFH, and a 12-​h infusion of abciximab or full-​dose tenecteplase with weight-​adjusted UFH for 48 h.  There was a reduction in the composite primary endpoints of 30-​day mortality, in-​hospital reinfarction, or in-​hospital refractory ischaemia in the enoxaparin and abciximab groups compared to the UFH group: 11.4% versus 15.4% (P = 0.0002) for enoxaparin, and 11.1% versus 15.4% (P < 0.0001) for abciximab. Similar to GUSTO V, the combination of abciximab and half-​dose tenecteplase did not have a mortality benefit compared to full-​dose tenecteplase alone but did result in significantly reduced in-​hospital infarction and refractory ischaemia. The need for urgent PCI was reduced in the GP IIb/​IIIa inhibitor and fibrinolytic combination group in both trials. In ASSENT-​3, major bleeding (other than intracranial haemorrhage, which was the same in both groups) was increased from 2.2% to 4.3% (P < 0.005). In those over the age of 75 years the risk of major bleeds was the greatest, with a threefold increase compared to younger patients. Facilitated PCI refers to pharmacological reperfusion treatment delivered prior to a planned PCI to bridge the PCI-​related time delay (69). Full-​dose fibrinolytic alone (ASSENT-​4 [70]), combinations of GP IIb/​IIIa inhibitor and low-​dose fibrinolytic (FINESSE [71]), or GP IIb/​IIIa alone (On-​TIME-​2 [72]) have been tested for this indication. In the Facilitated Intervention with Enhanced Reperfusion Speed to Stop Events (FINESSE) trial (71), primary PCI for patients with STEMI was preceded by early treatment with abciximab plus half-​dose reteplase or with abciximab alone and compared with abciximab administered immediately before the procedure. A total of 2452 patients was randomized into the study and, although there were improved markers of reperfusion as suggested by early ST-​segment resolution with the combination therapy, neither facilitation of PCI with reteplase plus abciximab nor facilitation with abciximab alone significantly improved the clinical outcomes, as compared with abciximab given at the time of PCI. Excess major bleeding was seen in the combination-​ facilitated group (14.5%) and abciximab-​facilitated group (10.1%) compared to the primary PCI group (6.9%; P < 0.0001). The trial was troubled with slow recruitment that resulted in it being terminated early as well as a low event rate that left it potentially

389

390

390

Section 5  

adjunctive drug therapies in percutaneous coronary intervention

underpowered. That said, the excess risk of bleeding (25 of 1000 patients treated) compared to the reduction in ischaemic composite endpoint (9 of 1000 patients treated) suggests that there is little role for facilitated PCI in this form, especially given recent data showing the strong relationship of in-​hospital bleeding with long-​term mortality (73).

Glycoprotein IIb/​IIIa inhibitors as adjunctive therapy to primary percutaneous coronary intervention In the context of primary PCI for the treatment of acute MI there are many reasons to suppose why additional antiplatelet therapy may be beneficial. The patient is in a prothrombotic state with activated platelets and inflammatory responses predisposing to potent platelet aggregation and further clot formation. Several randomized trials have assessed the value of abciximab given as adjunct therapy to primary PCI. The RAPPORT study (74) examined the benefit of abciximab in a placebo-​controlled trial in patients receiving primary angioplasty alone without stenting. They found a substantial reduction in the acute (30-​day) combined endpoint of death, reinfarction, and urgent TVR. However, the bleeding rates were excessive, and the 6-​month primary endpoint, which included elective revascularization, was not favourably affected. In the stenting era, the Abciximab Before Direct Angioplasty and Stenting in MI Regarding Acute and Long-​Term Follow-​Up (ADMIRAL) study (75) randomly assigned 300 patients with acute MI in a double-​ blind fashion either to abciximab plus stenting or placebo plus stenting. At 30  days, the primary endpoint, a composite of death, reinfarction, or urgent TVR, had occurred in 6.0% of the patients in the abciximab group, as compared with 14.6% of those in the placebo group (P = 0.01); at 6  months, the corresponding figures were 7.4% and 15.9% (P = 0.02). The better clinical outcomes in the abciximab group were related to the greater frequency of TIMI grade 3 coronary flow (according to the classification of the Thrombolysis in Myocardial Infarction trial [76]) in this group than in the placebo group before the procedure (16.8% versus 5.4%; P = 0.01), immediately afterward (95.1% versus 86.7%; P = 0.04), and 6 months afterward (94.3% versus 82.8%; P = 0.04). One major bleeding event occurred in the abciximab group (0.7%); none occurred in the placebo group. The clinical benefits initially observed at 30 days were still apparent at 3-​year follow-​up (77). The CADILLAC trial (78) examined 2082 patients and also found a significant reduction in 6-​month composite endpoints in the abciximab plus stenting group compared to stenting alone (10.2% versus 11.5%; P < 0.001). Like ADMIRAL, the difference in the incidence of the primary endpoint was due almost entirely to differences in the rates of TVR. De Luca et al. (79) undertook a systematic review of 11 trials examining the use of abciximab in STEMI involving 27,115 patients. When compared with the control group, abciximab was associated with a significant reduction in short-​term (30  days) mortality (2.4% versus 3.4;P = 0.047) and long-​term (6–​12 months) mortality (4.4% versus 6.2%; P = 0.01) in patients undergoing primary PCI. Abciximab was also associated with a significant reduction in 30-​day reinfarction (1.0% versus 1.9%; P = 0.03) in the primary PCI group. Abciximab did not result in an increased risk of intracranial bleeding nor major bleeding complications§. §

In the STEMI patients examined given abciximab and fibrinolysis there was a significant increase in major bleeding (5.2% versus 3.1%; P48 h) was randomized. The incidence of periprocedural myonecrosis was non-​inferior in the 5 g/​ dl or a decrease in haematocrit of >15%. Minor bleeding: spontaneous gross haematuria, or haematemesis, a decrease in haemoglobin >3 g/​dl with observed bleeding, decrease in haemoglobin >4 g/​dl with no observed bleeding, and all other observed bleeding.

39

Chapter 25 

IIIa trials examined of 2.5% compared to 1.3% in the placebo control (P < 0.0001) (55). Intracranial haemorrhage was a rare complication, occurring in only 0.08% of patients. GP IIb/​IIIa inhibitors were not associated with a significantly higher rate of intracranial haemorrhage or total stroke. Safety results were similar in patients treated with and without heparin. There is some concern about increased bleeding risk in patients receiving GP IIb/​IIIa inhibitors with chronic renal failure, although studies have yielded conflicting results. Best et al. found no increase in bleeding complications in patients receiving abciximab with renal impairment (112), but other studies have suggested increased risk (113, 114). Caution is advised, especially when patients have received clopidogrel loading and enoxaparin, which have been associated with a further increment in bleeding risk (115).

Thrombocytopaenia Thrombocytopaenia (less than 100×109/​l) is an uncommon, rarely fatal, but nevertheless concerning complication of GP IIb/​IIIa inhibitor usage. The incidence across the clinical trials is approximately 5% of patients involving all agents, although the incidence is greater in patients receiving abciximab than the small-​molecule agents. The precise mechanism of thrombocytopaenia is not well understood but it is possibly immune-​mediated given the increased prevalence of acute severe*, and profound** thrombocytopaenia on readministration of abciximab (2.8% and 2%, respectively) compared to the incidence after first-​time administration of 1% for severe and 0.4% for profound observed (30). Patients who form human antichimeric antibodies following abciximab administration (approximately 6%) are more prone to develop severe thrombocytopaenia on subsequent occasions (29). Because heparin is often administered concomitantly with GP IIb/​ IIIa inhibitors, heparin-​ induced thrombocytopaenia (HIT) must be excluded. HIT tends to occur 5–​10  days after the heparin therapy has been initiated rather than a precipitous reduction in platelet count within 24 h of initiation of the GP IIb/​IIIa inhibitor. Following HIT, immunoglobulin G antibodies against platelet factor 4 complexes may be detectable for 4–​6 weeks after the episode. If HIT and other causes of platelet consumption (such as disseminated intravascular coagulopathy) have been excluded, GP IIb/​ IIIa inhibitor-​ induced thrombocytopaenia should be treated by the immediate discontinuation of the agent. In the case of abciximab, platelet transfusion can be given if there is evidence of clinical bleeding, but not if small-​molecule agents have been administered. In this case, because of their weak binding and short duration of action, platelet counts will normally return to normal within 12–​124 h after the cessation of the drug.

Other adverse reactions Side effects of nausea, vomiting, hypotension, bradycardia, chest pain, back pain, headache, fever, puncture site pain, and, rarely, cardiac tamponade, adult respiratory distress, and hypersensitivity reactions have all been reported with GP IIb/​IIIa inhibitors (29, 31, 39; Table 25.2.

* Platelet count drop to less than 50×109/​l within 24 h of infusion. ** Platelet count drop to less than 20×109/​l within 24 h of infusion.

current status of glycoprotein iib/iiia inhibitors

Table 25.2  Contraindications and cautions relating to the use of glycoprotein (GP) IIb/​IIIa inhibitors Contraindications to use of GP IIb/​ Caution to use IIIa inhibitors (125) ◆

Active internal bleeding Major surgery ◆ Intracranial or intraspinal surgery or trauma within last 2 months ◆ Stroke within last 2 years ◆ Intracranial neoplasm ◆ Arteriovenous malformation or aneurysm ◆ Severe uncontrolled hypertension ◆ Haemorrhagic diathesis ◆ Thrombocytopaenia ◆ Vasculitis ◆ Hypertensive retinopathy ◆ Breastfeeding ◆ Severe hepatic impairment ◆

◆

Concomitant use of drugs that increase risk of bleeding ◆ Discontinue if uncontrollable serious bleeding occurs or emergency cardiac surgery needed ◆ Elderly ◆ Hepatic impairment ◆ Renal impairment ◆ Pregnancy ◆ Measure baseline prothrombin time, activated clotting time, activated partial thromboplastin time, platelet count, haemoglobin, and haematocrit ◆ Monitor haemoglobin and haematocrit 12 h and 24 h after start of treatment and platelet count 2–​4 h and 24 h after start of treatment

Other use of glycoprotein IIb/​IIIa inhibitors Stroke Two trials involving 474 patients were examined in a recent Cochrane review on the use of GP IIb/​IIIa inhibitors in acute ischaemic stroke (116). Only data for 414 patients treated within 6 h were considered. Patients were treated with intravenous abciximab or placebo. Treatment with abciximab was associated with a non-​ significant reduction of death and dependency combined (OR 0.79; 95% CI 0.54–​1.17) and of death alone (OR 0.67; 95% CI 0.36–​1.25). Treatment with abciximab was associated with a non-​ significant increase of symptomatic intracranial haemorrhages (OR 4.13; 95% CI 0.86–​19.67) and of major extracranial haemorrhages (OR 1.51; 95% CI 0.25–​9.12). A Cochrane review examined four trials involving 1365 participants and concluded that, based on the available evidence, for individuals with acute ischaemic stroke, GPIs were associated with a significant risk of intracranial haemorrhage with no evidence of any reduction in death or disability in survivors. They do not support their routine use in clinical practice. The conclusion was driven by trials of abciximab, which contributed 89% of the total number of study participants considered (117).

Peripheral vascular disease There has been sparse data examining the benefit of adjunctive antiplatelet therapy with GP IIb/​IIIa inhibitors in peripheral percutaneous interventions, in particular for critical limb ischaemia and peripheral vascular disease (118). To date, trials have been small, single-​centre studies, but initial results have been encouraging, showing a reduction in need for reintervention and improved angiographic patency at 6 months (119, 120).

393

394

394 Section 5  

Trial

Year

Study population

Drug

Groups**

Placebo Number

GP IIb/​IIIa inhibitor %

ARR (%)

RR

95% CI

P value

Notes

Number

%

10.3

49/​708

6.9*

3.4

0.68

0.47–​0.95 0.022

↑­bleeding rates due to high dose UH

PCI trials EPIC (29)

1994

High-​risk PCI; including ACS Abciximab

bolus; bolus + infusion

EPILOG (31)

1997

All PCI; ACS excluded

Abciximab

+standard or low 85/​939 dose UH

9.1

35/​935

3.7*

5.4

0.41

0.28–​0.61 70% (angio) >70% (USS)

Med Rx:CAS Med Rx:CEA

–​

–​

4 years

Stroke/​death

5 years

All except ACT-​1 have not yet been published. CAS, Carotid artery stenting; CEA, carotid endarterectomy; CVA, cerebrovascular accident; ns, not significant; MI, myocardial infarction; ns, not significant; USS, ultrasound scan.

689

Chapter 46 

carotid artery stenting

SPACE-​2 (ISRCTN 78592017) trial and CREST-​2 (NCT02089217) will allow only indirect comparisons between CAS and CEA in asymptomatic patients (95–​99). For the moment, an improvement in medical therapy and thus stroke risk suggests that any intervention has to be extremely safe and effective to be used routinely for primary prevention (Fig 46.4).

jeopardized by restenosis, perhaps because of poor collaterals in the circle of Willis, then repeat intervention may be more justified. Both balloon angioplasty with or without stent (106, 107), and CEA (108, 109) have been used, but the CEA is more challenging as the original stent has to be removed. The use of drug-​ eluting balloons, drug-​eluting stents, and even a bioabsorbable scaffold have been described (110–​112).

The restenotic lesion—​after carotid artery stenting

Key investigations

Stents reduce restenosis compared to balloon angioplasty in the ICA. In the CAVATAS trial, out to 8  years follow-​up, restenosis (>70%) for CEA was 14%, for CAS was 17%, but with angioplasty without stent was 34% (100). The rate of significant (>80% restenosis) has been estimated at 6.4% at 5 years of follow-​up in registry data (101, 102). The CAVATAS data suggests that restenosis >50% does carry some risk—​the rate of ipsilateral vascular events in follow-​up was 10.7% if there was no restenosis versus 22.7% with restenosis (P = 0.02). Restenosis occurs more often if the stent was placed for a restenotic lesion and in patients with diabetes (103). Progression to occlusion in in-​stent restenosis is more likely if the restenosis is diffuse and extends outside of the stent. It has been suggested that 6-​monthly duplex scans be utilized for follow-​up after CAS (103). Detection of restenosis remains an issue. Placing a stent alters the fluid dynamics in the carotid artery, making velocity criteria for stenosis that exist for native ICA stenosis invalid. The stent struts make 2D imaging difficult. Angiographic validation studies have suggested very variable cut-​ off recommended for peak systolic velocities in the ICA (104, 105). Suspicion of high-​grade stenosis should lead to computed tomography (CT) angiogram confirmation, or invasive angiography, although, of course, this carries a risk of complications. For symptomatic stenosis, repeat intervention seems appropriate as for restenosis post CEA and native ICA disease. For asymptomatic restenosis, data to confirm the benefits of reintervention are lacking. If cerebral blood supply is felt to be

6

Annual rates of stroke (%)

5

62

ACAS ‘any stroke’ 59

60

61

3 60

2

Because of the ease of access, the carotid duplex ultrasound scan has become the screening tool of choice (Fig 46.2). Two meta-​analyses have suggested that its sensitivity at detecting an ICA stenosis of 70–​99% on invasive angiography was 86–​90% with specificity 87–​ 94% (113, 114). There is considerable skill involved in performing and interpreting the scan, however, and trust has to be developed with the ultrasonographer. Repeat scans at different institutions can produce widely varying results (114). It should include 2D measurements, but also velocity measurements in assessing stenosis in the carotid bulb and both higher and lower in the carotid circulation. In addition, flow in the vertebral artery is assessed. Attempts have been made to try to improve plaque characterization, and these may develop further with the advent of virtual histology. There is some data to suggest that plaque morphology predicts outcome after intervention, although greater experience and use of proximal protection devices might significantly diminish this risk (115, 116).

Invasive angiography Selective angiography for assessing ICA stenosis carries a risk of TIA and CVA, and so is now rarely performed except in those with inconclusive non-​invasive results (Fig 46.5). The risks are higher in those with highly significant (>90%) stenoses and permanent (>7  days) neurological deficit can occur in 2.5% of cases (117). Using arch aortography and digital subtraction and intravenous contrast, non-​selective images can be obtained to allow

Any stroke; 50–99% stenosis Any stroke; 60–99% stenosis Any stroke; 70–99% stenosis

58

4

Carotid duplex scanning

61

Ipsilateral stroke; 50–99% stenosis Ipsilateral stroke; 60–99% stenosis Ipsilateral stroke; 70–99% stenosis

ACAS3 67

63 63

64

66

ACST 1–5 years4

ACAS3 65

ACAS ‘ipsilateral stroke’

66

ACST 1–5 years4

1 0 1984

77 44 44

45 54

ACST 6–10 years5 77

68 68

ACST 6–10 years5 1989

1994

1999

2004

54 55 69

2009

Year

Figure 46.4  Reduction in annual stroke rate over time (reference numbers reflect original paper). Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Cardiology. Naylor AR. Time to rethink management strategies in asymptomatic carotid artery disease, (9)116–​24. Copyright (2011).

689

690

690

Section 8  

non-coronary percutaneous interventions A

B

Figure 46.5  Invasive carotid angiography with digital subtraction. Tight restenosis after carotid endarterectomy shown on digital subtraction angiography (A). Post-​ stenting (clear arrow), the flow is improved (B). The external carotid artery is already occluded and not seen in either panel. The guide catheter is seen at the lower end of the panels (solid arrow) and the distal protection device is in position (magnified insert).

measurement of stenosis. However, these methods have not been widely validated (118). With modern computed tomography angiography (CTA), the use of invasive angiography should really be periprocedural.

Magnetic resonance angiography Access to magnetic resonance angiography (MRA) is limited in the UK, but can produce excellent images of the aortic arch, great vessels, and intracerebral vasculature. Decisions on treatment may be swayed by the presence or absence of intracerebral collaterals from the circle of Willis. However, pseudostenoses can result from movement artifact. If combined with MRI complex brain imaging, this may become the best investigation after carotid duplex screening (Fig 46.6).

Computed tomography angiography With the advent of multislice CT, detailed imaging of the arch and great vessels can be achieved (Figs 46.7 and 46.8). In addition, the intracerebral circulation can be imaged. Invasive angiography remains the gold standard, but image quality is rapidly improving, allowing accurate assessment of the stenosis, as well as plaque composition; spatial resolution of 0.4 mm is now possible. Sensitivity for a 70–​99% ICA stenosis is rated at 85% and specificity at 92% (119). The main limitation of CT is that brain imaging is, as yet, not as detailed as with MRI. However, multimodality CT provides substantial benefits in stroke patients, being faster and more accessible than MRI (120). As for invasive angiography, the percentage stenosis needs to be calculated with reference to the distal diameter of the ICA (NASCET measurement), although, given that the cross-​section of the ICA stenosis is available, the equivalent to the duplex stenosis (ECST measurement) can also be assessed (Fig 46.2).

Summary of imaging techniques The ever-​increasing number of methods for assessment of carotid stenosis led to controversy about the gold standard. Good-​ quality comparative studies are lacking. As the major trials have used duplex scanning or intra-​arterial invasive angiography, new techniques will need to be validated against these until they too have been used to independently select patients for intervention with proven improvement in outcome. Units will vary as to which technique to rely upon, but duplex scanning will remain the best method of initial assessment, being non-​invasive, repeatable, and hopefully reproducible. Knowledge of the anatomy and lesion type assist in making the final decision between stenting and surgery as both are thought to be associated with elevated risk during CAS. CT scanning can provide useful information on extracranial and intracranial pathology to aid operation planning.

Technique for carotid artery stenting Planning and pre-​treatment Once the patient has been reviewed at the multidisciplinary team (MDT) meeting, and accepted for stenting, the imaging should be reviewed again by the interventionalist performing the case. Knowledge of the anatomy of the peripheral access vessels (usually femoral), arch and cervical vessels allows appropriate planning. With the advent of simulators allowing case scenarios, a virtual procedure might be performed in high-​risk cases to assess equipment choices (121).

Preoperative medication The patient should be adequately hydrated, as with all vascular catheterization procedures, and renal protection given if there is known renal dysfunction. Use of N-​acetyl-​cysteine (NAC) is no

691

Chapter 46 

carotid artery stenting

Figure 46.6  Magnetic resonance angiography of carotid artery. A right internal carotid artery stenosis is clearly seen (arrow).

longer recommended, but still often used, usually given at 600 mg bd starting the day before the procedure (122). Aspirin and clopidogrel need to be initiated at least 24 h prior to the procedure (123). A loading dose of 300 mg of each is given if the patient is not on them regularly, with 75 od thereafter. Antihypertensive medication is reviewed, and the beta-​blocker stopped for the day of the procedure. The aim is to have a systolic pressure between 120–​180 mmHg at the start of the procedure.

Perioperative medication Unfractionated heparin (UFH) is administered prior to attempting to cannulate the common carotid artery (CCA) at a dose of 70–​ 100 U/​kg. The activated clotting time (ACT) should be maintained at 250–​300 s. If the procedure is prolonged for >1 h, then the ACT should be rechecked and more heparin given if required. Although data is limited, if heparin cannot be tolerated, bivalirudin (Angiomax, The Medicines Company, USA) provides an alternative. A bolus of 0.75 mg/​kg is followed by 1.75 mg/​kg/​h infusion for the duration of the procedure (124). Use of adjunctive pharmacology has improved outcomes in PCI. Data in CAS suggests that use of DPDs gives greater benefit than use of glycoprotein IIb/​IIIa agents, and, in large series, no benefit from these agents has yet been shown (125). There is a danger of haemorrhagic transformation if embolic stroke occurs with these agents on board. However, they may find a place in addition to DPDs in reducing platelet aggregation within the filter device, and when an obvious thrombotic complication has occurred (126, 127). There is no data as yet on stronger antiplatelets such as ticagrelor or prasugrel in this setting.

Gaining a stable working platform The patient should be comfortable with a head cushion to maintain a stable position. There should be electrocardiographic and

Figure 46.7  3D computed tomography angiogram of a right internal carotid artery stenosis (arrow). Volume rendering allows a 3D appreciation of the anatomy.

haemodynamic monitoring attached as bradycardia and hypotension are not uncommon during the procedure. An intravenous line should be available and functional. In some laboratories, the patient is given a noise-​making device in the contralateral hand to the ICA being treated to provide an audio clue to neurological decline during the procedure. This is usually supplemented by communication with the patient during the case to assess speech and cognitive function, and sedation is thus kept to a minimum. A standard technique should be employed to gain familiarity with the approach. Although radial access CAS was first described in 1999, and has been suggested as feasible, the femoral approach—​to increase catheter support and allow larger stents to be placed—​remains the standard approach (128). Crossover from radial to femoral is more common than for coronary angiography, and there appears to be no gain in terms of reduced complication rates (129). Direct cervical access to the CCA with subsequent percutaneous completion has also been tried, but requires surgical expertise (130). All techniques involve gaining femoral access with a Seldinger technique and a 6 F to 8 F sheath. A 5 F pigtail catheter is used to perform a non-​selective arch aortogram if this has not been done as part of the work-​up. Prescreening of the CT scan will allow an optimal aortogram to be performed. Once the origin of the great arteries has been visualized, a 5 F diagnostic glide catheter is used to gain access to the innominate artery (if treating the right ICA) or left CCA in the left anterior oblique (LAO) position (Fig 46.5).

691

692

692

Section 8  

non-coronary percutaneous interventions

Figure 46.8  Computed tomography angiogram measurements. Measurement of the size of the carotid artery on computed tomography to assess dimensions for stenting. A more ectatic segment of carotid artery is seen at the carotid bulb (arrow).

Problems arise with a bovine origin of the left CCA (take-​off from innominate artery) or origin of the innominate artery from the earlier part of the ascending aorta (a type III arch). The introduction of the Magellan endovascular robot (Hansen Medical, USA) has started a new concept in vascular intervention. The system consists of control panel with keyboard and joystick.

This is connected to the robotic arm at the end of the angio table. The arm controls the movement of the robotic wire and catheter. The potential advantages of this system include predictability, accuracy, efficiency of movement, as well as radiation protection (the operator is outside the room). The initial preclinical work on cannulation of supra-​aortic vessels indicates significant reduction

693

Chapter 46 

in wall hits and shortening of procedure time when compared to manual catheter cannulation. The bench top tests also suggest that the robotic catheter would produce less contact force with the arterial wall and more catheter stability. Furthermore, initial clinical data demonstrates that robotic manipulation around the aortic arch produces less embolic hits, as measured by transcranial Doppler, than manual catheter manipulation (131). More clinical data and well designed randomized trials are needed to fully understand and estimate the benefits and the role of robotics in carotid intervention (Fig 46.9). Once the CCA is intubated, images are taken of the carotid stenosis in two planes to define the severity of the stenosis and the best image to separate the origin of the external carotid artery (ECA). This might require a lateral, LAO, anteroposterior, or right anterior oblique (RAO) projection. In addition, lateral and anteroposterior projections of the intracerebral vessels are taken to compare run-​off at the end in case of complications. Digital subtraction angiography is ideal, but not essential. A 0.035'' 190-​cm Terumo glidewire is used to help advance the 5-​F diagnostic catheter into the ECA. This is made easier if roadmapping is available. Care needs to be taken to avoid inadvertently crossing the ICA stenosis with the Terumo wire. The 5-​F catheter is then advanced into the ECA, and the Terumo wire exchanged for an 0.035'' 300-​cm stiff Amplatzer wire. This allows the support needed to pass either a sheath or a guide catheter into the CCA. Most commonly the procedure in the hands of radiologists involves a long sheath as the conduit, whilst cardiologists prefer a guide catheter. Care needs to be taken not to perforate a branch of

carotid artery stenting

the ECA with the stiff wire. The stiff wire is then withdrawn and an image or roadmap of the lesion taken to allow easy passage of the angioplasty wire or embolic protection device (DPD). An alternative to a guide catheter is to use a shuttle sheath, which may be preloaded on the diagnostic 5-​F glide catheter. Once entry into the ECA is achieved with the glide catheter, the 6 F or 7 F 90-​cm sheath is again positioned in the CCA below the carotid bifurcation, taking care not to disturb the ICA stenosis with the introducer. The image intensifier is now positioned to image the lesion and the ICA at least to the base of the skull and allow an embolic protection device to be placed.

Protection or not? Microembolization is thought to be the major cause of perioperative neurological complications during CAS (132). Transcranial Doppler measurements suggest that, without distal protection, the incidence of microembolization is higher with CAS than CEA (133). MRI scanning shows more white matter lesions than with CEA (86). Thus prevention of distal embolization would seem logical. This can be achieved with distal filter devices, distal occlusion devices, or proximal occlusion devices. (For a full review of this subject see Boisiers et al. [134]) (Table 46.9). Most cardiologists are likely to prefer distal protection filters owing to their familiarity with them. Large-​scale trials comparing outcomes with and without distal protection are lacking, but registry data suggest that some form of protection device should be used in most cases. Event rates

Figure 46.9  Robotic carotid artery stenting. The Magellan robotic system reduces the risks of endovascular procedures. A robotic arm controls catheter position inside the body. Image courtesy of Hansen Medical.

693

694

694

Section 8  

non-coronary percutaneous interventions

Table 46.9  Embolic protection devices used for carotid artery stent procedures with advantages and disadvantages listed Technique

Advantages

Disadvantages

Balloon occlusion ◆ Percusurge (Medtronic) ◆ Twin-​one (Minvasys)

◆

◆

Can be used in patients with low cardiac output ◆ No embolic debris if used correctly

◆ ◆ ◆ ◆

Distal Filter ◆ Angioguard (Cordis) ◆ Accunet (Guidant) ◆ EZ Filter (Boston) ◆ Interceptor (Medtronic) ◆ NeuroShield (Abbott) ◆ Spider Rx (Ev3) ◆ FiberNet (Lumen Biomedical)

◆ Maintain

◆

anterograde perfusion ◆ Allow imaging whilst in use

◆

Proximal occlusion ◆ Parodi (ArteriA) ◆ MOMA (Invatec) ◆ ENROUTE (Silk Road Medical)

◆

◆ ◆ ◆

Can be used in large vessels ◆ No need to cross ICA lesion prior to protection ◆ No embolic debris if used correctly

Requires lesion to be crossed without protection to deliver device Maximum vessel size 6 mm Can lead to cerebral hypoperfusion No flow of contrast, so imaging difficult Risk of trauma to ICA with movement of the filter Requires lesion to be crossed without protection to deliver device (device crossing profile up to 3.9 F) Filter may clog or get overloaded Can be difficult to retrieve in tortuous vessels Maximum size 8 mm Risk of trauma to ICA with movement of the filter

◆

Can lead to cerebral hypoperfusion ◆ Large sheath size in femoral artery (10 F or 11 F) ◆ Requires disease-​free ECA

ICA, Internal carotid artery; ECA, external carotid artery.

(30-​day stroke and death) in symptomatic patients having CAS in the pre-​protection era were 6.7% versus 2.82% in the modern age (135). For asymptomatic patients, rates were 3.97% and are now 1.75%, respectively. Distal balloon occlusion requires crossing the ICA lesion with the device, inflation of a low-​pressure balloon to block the passage of debris distally during CAS, and removal of debris by aspiration in the ICA prior to the end of the procedure. Because bloodflow is interrupted, the patient may become restless because of cerebral hypoperfusion if collateral circulation in the brain is not adequate. If hypoperfusion occurs, the balloon can be temporarily deflated, but of course that risks distal embolization. Therefore, the CAS procedure must be completed within a few minutes. An additional problem with this technique is that angiographic images cannot be taken during the procedure as dye will preferentially pass into the ECA during balloon occlusion of the ICA. Filter DPDs use 80–​100-​micron filters inserted via the guide catheter or sheath, and deployed beyond the ICA stenosis. Several types exist, some mounted on a wire, whilst others pass over a guide wire

that has already crossed the lesion. The bloodflow is maintained, and so the procedure is better tolerated. In addition, repeated contrast images can be obtained. However, movement of the device can result in spasm or even dissection of the ICA, and, if apposition around the circumference of the vessel wall is not perfect, debris can still pass upwards to the brain. The device is recaptured after the CAS procedure and withdrawn through the stent to remove the debris that has been caught. These devices would be favoured if pre-​ procedure assessment suggested an incomplete circle of Willis, and hence poor collateral circulation. The third device involves the proximal occlusion of the CCA and the ECA (136, 137). A large catheter with an extension arm for the ECA is advanced into the CCA. Occlusion balloons mounted on the catheter are inflated in the CCA and then the ECA. This creates a negative pressure in the ICA and should allow the collaterals in the circle of Willis to produce reverse flow or at least no forward flow, thus preventing distal embolization into the ICA. The problems highlighted with distal balloon occlusion, about cerebral perfusion and imaging, also exist with this technique, but the lesion does not need to be crossed, making it useful if it is felt to be particularly thrombus-​laden or high risk. Once the stent procedure is completed, aspiration is performed to clear debris from the ICA prior to device removal. These devices are larger and so not suitable for patients with peripheral vascular disease. Some data suggest smaller, less numerous, MRI lesions with such devices, but there are no data on hard neurological outcomes (137, 138). The ENROUTE Transcarotid Neuroprotection System (NPS; Silk Road Medical Inc, Sunnyvale, CA, USA) is a hybrid protection system, which involves a small cut-​down over the CCA at the root of the neck and percutaneous access of the common femoral vein in the groin. The closed circuit system allows high-​or low-​flow reversal or surgical cessation of arterial flow in the carotid artery, using a surgical clamp or sling. The proposed benefits of this system include the easy access of the artery regardless of the arch anatomy as well as safe access of the lesion with flow reversal. The system has an incorporated filter that captures debris as small as 200 microns before re-​ entering the venous system. The ROADSTAR multicentre trial demonstrated safety and efficacy of this system with a low overall stroke risk of 1.4%. However, there was a record of cranial nerve injury in this controlled trial, reflecting the relative invasiveness of this system (139). Possible complications with the use of DPDs include difficulty in deploying the device, effectiveness of embolus capture, vessel injury by the device, and problems with retrieval (140). Although one small study using MRI scanning has suggested less debris passes to the brain with proximal occlusion devices, patient factors, operator experience, and perhaps the choice of stent are likely to be more relevant (141). In summary, data on whether to use devices is not yet categorical, although registry data favour their use. The choice of device, if any, needs to be made on the anatomy of the lesion and operator experience with the equipment. Although there are no randomized controlled trials, the published data suggest the superiority of proximal balloon occlusion filters over other protection methods (137).

Balloon and stent Once the DPD is in place, if the lesion is tight, with less than a 2-​ mm diameter, or looks calcified, pre-​dilatation with a 3 × 20 mm balloon is recommended. Rapid exchange systems are usually

695

Chapter 46 

preferred to over-​the-​wire systems for ease of use and wire stability. Direct stenting may be performed for less severe stenoses. It is vital to give atropine 600 micrograms prior to balloon inflation. Failure to do so will result in bradycardia and hypotension, which can be difficult to manage and may be prolonged.

Which stent? The choice of stent depends on the characteristics of the lesion and the shape of the vessel. Self-​expanding stents have an inherent ability to expand, minimizing the concern about stent collapse if carotid palpation is performed after stenting. It is important to be familiar with a few stents and how they perform rather than to use a large range. Most important is to ensure the stent is large enough to expand into the ICA and, as the lesion usually extends into the CCA, the CCA also. Most stents shorten as they expand, and so adequate length of stent is important. Unlike the coronary, stent length has not been linked to restenosis, and so a 20–​40-​mm length is often taken to avoid geographical miss. The stent can be positioned using bony markers, a reference image on a second screen, or using roadmapping. Once the stent is in place, post-​dilatation is usually required to allow the stenosis to be reduced to 50 patients was linked to reduced risk, whilst larger annual numbers also suggested improved outcomes. Thus, maintaining practice is vital (160, 161). What is also clear in our modern practice with full duty of candour is that, if the personal level of experience of the operator might be described as ‘on the learning curve’, then this is likely to be disclosable as part of the consent process. Failure to do so could leave the operator open to litigation in the event of complications (162).

Conclusion CEA is not as invasive as CABG surgery, whilst CAS carries higher risks than PCI since the brain is more unforgiving than the myocardium. However, the move towards the development of less invasive treatments is unrelenting and, with the appropriate training, many interventionalists will be able to perform CAS safely. It is important to pay attention to initial patient selection and appropriate investigation. This is best done in the context of an MDT. Careful selection of technique to be employed, choice of equipment, and meticulous preoperative, perioperative, and postoperative care is vital to ensure that the procedure is performed at lowest possible risk. Careful liaison with the MDT allows clinical governance and audit to be transparent and honest. The evidence base comparing CAS to CEA is enlarging, with both registry and trial data. It is clear from long-​term data of comparative trials that, if one can avoid a complication in the first 30 days, both treatments provide durable results. Thus, the real question is, in any institution, which treatment can offer the lowest operative risk. This should be on the basis of internal, audited data, ideally involving neurological assessment, and not self-​reporting of outcomes. In the symptomatic patient, the current consensus regarding the timing of carotid intervention suggests that intervening in the

697

698

698

Section 8  

non-coronary percutaneous interventions

first 48 h does not increase the stroke/​death risks and it is safe in the hyperacute phase with CEA. Even if the operative event rate is higher, the net reduction in stroke appears to favour early intervention, rather than letting ‘nature take its course’ and delaying intervention. CEA remains superior to CAS at this stage, except perhaps in high-​volume centres with proven excellence. For patients over 75, the original surgical trial suggested high benefit for CEA, whilst CAS registries and trials caution against CAS in those over 80. In those over 80, there has to be a careful assessment of operative risk due to comorbidities and operative benefit with CEA (or CAS) before deciding on any therapy beyond best medical management. Indeed, in any patient, an MDT assisted by a scoring system, such as Delphi, would help avoid patient harm (163). The high-​risk patient could be offered CAS if there is an established program. A more significant question with modern medical management of atherosclerotic disease is if any intervention, CEA or CAS, is needed in a patient with asymptomatic but significant known carotid stenosis. CAS is safe in this setting, but medical therapy may well have reduced the risk of stroke so much already that intervention, much as in stable coronary disease, may not produce any prognostic advantage.

References 1. Kung HC, Hoyert DL, Xu J, Murphy SL. Deaths: final data for 2005. Natl Vital Stat Rep 2008;56(10):1–​120. 2. Thrift AG, Thayabaranathan T, Howard G, et al. Global stroke statistics. Int J Stroke 2017;12(1):13–32. 3. Adams HP, Jr., Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993;24(1):35–​41. 4. Al-​Lamee R, Davies J, Malik IS. What is the role of coronary angioplasty and stenting in stable angina? BMJ 2016;352:i205. doi: 10.1136/​bmj.i205 5. Stroke: Diagnosis and initial management of acute stroke and transient ischaemic attack (TIA). 2008. http://​www.nice.org.uk/​nicemedia/​pdf/​ CG68NICEGuideline.pdf. 6. Griessenauer CJ, Medin C, Maingard J, et al. Endovascular Mechanical Thrombectomy in Large-Vessel Occlusion Ischemic Stroke Presenting with Low National Institutes of Health Stroke Scale: Systematic Review and Meta-Analysis. World Neurosurg 2018;110:263–69. 7. Fischer CM, Gore I, Okabe N, White PD. Atherosclerosis of the carotid and vertebral arteries: extracranial and intracranial. J Neuropath Exp Neurol 1965;24:455–​76. 8. Sacco RL, Kasner SE, Broderick JP, et al. AHA/​ASA Expert Consensus Document: an updated evised definition of stroke for the 21st century. Stroke 2013;44:2064–​89. 9. Collaborative overview of randomised trials of antiplatelet therapy—​ I: Prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. Antiplatelet Trialists’ Collaboration [see comments] [published erratum appears in BMJ 1994;308(6943):1540]. BMJ 1994;308(6921):81–​106. 10. Vane JR, Meade TW. Second European Stroke Prevention Study (ESPS 2): clinical and pharmacological implications. J Neurol Sci 1997;145(2):123–​5. 11. Halkes PH, van Gijn J, Kappelle LJ, Koudstaal PJ, Algra A. Aspirin plus dipyridamole versus aspirin alone after cerebral ischaemia of arterial origin (ESPRIT): randomised controlled trial. Lancet 2006;367(9523):1665–​73. 12. Diener HC, Bogousslavsky J, Brass LM, et al. Aspirin and clopidogrel compared with clopidogrel alone after recent ischaemic stroke or transient ischaemic attack in high-​risk patients (MATCH): randomised, double-​blind, placebo-​controlled trial. Lancet 2004;364(9431):331–​7.

13. Sacco RL, Diener HC, Yusuf S, et al. Aspirin and extended-​release dipyridamole versus clopidogrel for recurrent stroke. N Engl J Med 2008;359(12):1238–​51. 14. Baigent C, Keech A, Kearney PM, et al. Efficacy and safety of cholesterol-​lowering treatment: prospective meta-​analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet 2005;366(9493):1267–​78. 15. The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Investigators. High-​dose atorvastatin after stroke or transient ischemic attack. N Engl J Med 2006;355(6):549–​59. 16. Eastcott HH, Pickering GW, Rob CG. Reconstruction of internal carotid artery in a patient with intermittent attacks of hemiplegia. Lancet 1954;267(6846):994–​6. 17. DeBakey ME. Successful carotid endarterectomy for cerebrovascular insufficiency. Nineteen-​year follow-​up. JAMA 1975;233(10):1083–​5. 18. Murie JA. Carotid endarterectomy in Great Britain and Ireland: practice between 1984 and 1992. Br J Surg 1994;81(6):827–​31. 19. DuBose J, Recinos G, Teixeira PG, Inaba K, Demetriades D. Endovascular stenting for the treatment of traumatic internal carotid injuries: expanding experience. J Trauma 2008;65(6):1561–​6. 20. Bockenheimer SA, Mathias K. Percutaneous transluminal angioplasty in arteriosclerotic internal carotid artery stenosis. AJNR Am J Neuroradiol 1983;4(3):791–​2. 21. Diethrich EB, Ndiaye M, Reid DB. Stenting in the carotid artery: initial experience in 110 patients. J Endovasc Surg 1996;3(1):42–​62. 22. Yadav JS, Roubin GS, Iyer S, et al. Elective stenting of the extracranial carotid arteries. Circulation 1997;95(2):376–​81. 23. Theron J, Raymond J, Casasco A, Courtheoux F. Percutaneous angioplasty of atherosclerotic and postsurgical stenosis of carotid arteries. AJNR Am J Neuroradiol 1987;8(3):495–​500. 24. International Carotid Stenting Study Investigators. Carotid artery stenting compared with endarterectomy in patients with symptomatic carotid stenosis (International Carotid Stenting Study): an interim analysis of a randomised controlled trial. Lancet 2010;375:985–​97. 25. Bonati LH, Lyrer P, Ederle J, Featherstone R, Brown MM. Percutaneous transluminal balloon angioplasty and stenting for carotid artery stenosis. Cochrane Database Syst Rev 2012;9CD000515. 26. Veith FJ, Amor M, Ohki T, et al. Current status of carotid bifurcation angioplasty and stenting based on a consensus of opinion leaders. J Vasc Surg 2001;33(2 Suppl.):S111–​16. 27. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-​grade carotid stenosis. N Engl J Med 1991;325(7):445–​53. 28. Mayberg MR, Wilson SE, Yatsu F, et al. Carotid endarterectomy and prevention of cerebral ischemia in symptomatic carotid stenosis. Veterans Affairs Cooperative Studies Program 309 Trialist Group. JAMA 1991;266(23):3289–​94. 29. European Carotid Surgery Trialists’ Collaborative Group. MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70–​99%) or with mild (0–​29%) carotid stenosis. Lancet 1991;337(8752):1235–​43. 30. Barnett HJ, Taylor DW, Eliasziw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators. N Engl J Med 1998;339(20):1415–​25. 31. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998;351(9113):1379–​87. 32. Rothwell PM, Eliasziw M, Gutnikov SA, et al. Analysis of pooled data from the randomised controlled trials of endarterectomy for symptomatic carotid stenosis. Lancet 2003;361(9352):107–​16. 33. Moore WS, Barnett HJ, Beebe HG, et al. Guidelines for carotid endarterectomy. A multidisciplinary consensus statement from the Ad Hoc Committee, American Heart Association. Circulation 1995;91(2):566–​79.

69

Chapter 46 

34. Barnett HJ, Gunton RW, Eliasziw M, et al. Causes and severity of ischemic stroke in patients with internal carotid artery stenosis. JAMA 2000;283(11):1429–​36. 35. Inzitari D, Eliasziw M, Gates P, et al. The causes and risk of stroke in patients with asymptomatic internal carotid artery stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators. N Engl J Med 2000;342(23):1693–​700. 36. Wolff T, Guirguis-​Blake J, Miller T, Gillespie M, Harris R. Screening for carotid artery stenosis: an update of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2007;147(12):860–​70. 37. Rockman CB, Jacobowitz GR, Gagne PJ, et al. Focused screening for occult carotid artery disease: patients with known heart disease are at high risk. J Vasc Surg 2004;39(1):44–​51. 38. Cina CS, Safar HA, Maggisano R, Bailey R, Clase CM. Prevalence and progression of internal carotid artery stenosis in patients with peripheral arterial occlusive disease. J Vasc Surg 2002;36(1):75–​82. 39. Thapar A1, Jenkins IH, Mehta A, Davies AH. Diagnosis and management of carotid atherosclerosis. BMJ 2013;346:f1485. doi: 10.1136/​bmj.f1485. 40. The CASANOVA Study Group. Carotid surgery versus medical therapy in asymptomatic carotid stenosis. Stroke 1991;22(10):1229–​35. 41. Hobson RW, Weiss DG, Fields WS, et al. Efficacy of carotid endarterectomy for asymptomatic carotid stenosis. The Veterans Affairs Cooperative Study Group. N Engl J Med 1993;328(4):221–​7. 42. Mayo Asymptomatic Carotid Endarterectomy Study Group. Results of a randomized controlled trial of carotid endarterectomy for asymptomatic carotid stenosis. Mayo Clin Proc 1992;67(6):513–​18. 43. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995;273(18):1421–​8. 44. Halliday A, Mansfield A, Marro J, et al. Prevention of disabling and fatal strokes by successful carotid endarterectomy in patients without recent neurological symptoms: randomised controlled trial. Lancet 2004;363(9420):1491–​502. 45. Rothwell PM, Gibson RJ, Slattery J, Sellar RJ, Warlow CP. Equivalence of measurements of carotid stenosis. A comparison of three methods on 1001 angiograms. European Carotid Surgery Trialists’ Collaborative Group. Stroke 1994;25(12):2435–​9. 46. Simes J. Long-​term effectiveness and safety of pravastatin in 9014 patients with coronary heart disease and average cholesterol concentrations: the LIPID trial follow-​up. Lancet 2002;359(9315):1379–​87. 47. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). [see comments]. Lancet 1996;348(9038):1329–​39. 48. Wennberg DE, Lucas FL, Birkmeyer JD, Bredenberg CE, Fisher ES. Variation in carotid endarterectomy mortality in the Medicare population: trial hospitals, volume, and patient characteristics. JAMA: the Journal of the American Medical Association 1998;279(16):1278–​81. 49. Theiss W, Hermanek P, Mathias K, et al. Pro-​CAS: a prospective registry of carotid angioplasty and stenting. Stroke 2004;35(9): 2134–​9. 50. Brener BJ, Brief DK, Alpert J, Goldenkranz RJ, Parsonnet V. The risk of stroke in patients with asymptomatic carotid stenosis undergoing cardiac surgery: a follow-​up study. J Vasc Surg 1987;5(2):269–​79. 51. Huh J, Wall MJ, Jr., Soltero ER. Treatment of combined coronary and carotid artery disease. Curr Opin Cardiol 2003;18(6):447–​53. 52. Naylor AR, Mehta Z, Rothwell PM, Bell PR. Carotid artery disease and stroke during coronary artery bypass: a critical review of the literature. Eur J Vasc Endovasc Surg 2002;23(4):283–​94. 53. Das SK, Brow TD, Pepper J. Continuing controversy in the management of concomitant coronary and carotid disease: an overview. Int J Cardiol 2000;74(1):47–​65.

carotid artery stenting

54. Durand DJ, Perler BA, Roseborough GS, et al. Mandatory versus selective preoperative carotid screening: a retrospective analysis. Ann Thorac Surg 2004;78(1):159–​66. 55. Fareed KR, Rothwell PM, Mehta Z, Naylor AR. Synchronous carotid endarterectomy and off-​pump coronary bypass: an updated, systematic review of early outcomes. Eur J Vasc Endovasc Surg 2009;37(4):375–​8. 56. Guzman LA, Costa MA, Angiolillo DJ, et al. A systematic review of outcomes in patients with staged carotid artery stenting and coronary artery bypass graft surgery. Stroke 2008;39(2):361–​5. 57. Goto T, Baba T, Yoshitake A, Shibata Y, Ura M, Sakata R. Craniocervical and aortic atherosclerosis as neurologic risk factors in coronary surgery. Ann Thorac Surg 2000;69(3):834–​40. 58. Illuminati G, Schneider F, Greco C, et al. Long-​term results of a randomized controlled trial analyzing the role of systematic pre-​ operative coronary angiography before elective carotid endarterectomy in patients with asymptomatic coronary artery disease. Eur J Vasc Endovasc Surg 2015;49(4):366–​74. 59. Lattimer CR, Burnand KG. Recurrent carotid stenosis after carotid endarterectomy. Br J Surg 1997;84(9):1206–​19. 60. Sterpetti AV, Schultz RD, Feldhaus RJ, et al. Natural history of recurrent carotid artery disease. Surg Gynecol Obstet 1989;168(3):217–​23. 61. Bartlett FF, Rapp JH, Goldstone J, Ehrenfeld WK, Stoney RJ. Recurrent carotid stenosis: operative strategy and late results. J Vasc Surg 1987;5(3):452–​6. 62. Healy DA, Zierler RE, Nicholls SC, et al. Long-​term follow-​up and clinical outcome of carotid restenosis. J Vasc Surg 1989;10(6):662–​8. 63. Moore WS, Kempczinski RF, Nelson JJ, Toole JF. Recurrent carotid stenosis: results of the asymptomatic carotid atherosclerosis study. Stroke 1998;29(10):2018–​25. 64. Stoney RJ, String ST. Recurrent carotid stenosis. Surgery 1976;80(6):705–​10. 65. Hobson RW, Goldstein JE, Jamil Z, et al. Carotid restenosis: operative and endovascular management. J Vasc Surg 1999;29(2):228–​35. 66. Endovascular versus surgical treatment in patients with carotid stenosis in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS): a randomised trial. Lancet 2001;357(9270):1729–​37. 67. Veith FJ, Amor M, Ohki T, et al. Current status of carotid bifurcation angioplasty and stenting based on a consensus of opinion leaders. J Vasc Surg 2001;33(2 Suppl):S111–​16. 68. McCarthy WJ, Wang R, Pearce WH, Flinn WR, Yao JS. Carotid endarterectomy with an occluded contralateral carotid artery. Am J Surg 1993;166(2):168–​71. 69. Rothwell PM, Slattery J, Warlow CP. Clinical and angiographic predictors of stroke and death from carotid endarterectomy: systematic review. BMJ 1997;315(7122):1571–​7. 70. Borger MA, Fremes SE, Weisel RD, et al. Coronary bypass and carotid endarterectomy: does a combined approach increase risk? A metaanalysis. Ann Thorac Surg 1999;68(1):14–​20. 71. Paciaroni M, Eliasziw M, Kappelle LJ, Finan JW, Ferguson GG, Barnett HJ. Medical complications associated with carotid endarterectomy. North American Symptomatic Carotid Endarterectomy Trial (NASCET). Stroke 1999;30(9):1759–​63. 72. Rerkasem K, Rothwell PM. Local versus general anaesthesia for carotid endarterectomy. Cochrane Database Syst Rev 2008;(4):CD000126. 73. Roubin GS, New G, Iyer SS, et al. Immediate and late clinical outcomes of carotid artery stenting in patients with symptomatic and asymptomatic carotid artery stenosis: a 5-​year prospective analysis. Circulation 2001;103(4):532–​7. 74. Alberts MJ. Results of a prospective multicentre randomised trial of carotid artery stenting vs carotid endarterectomy. Stroke 2001;32:325. 75. Brooks WH, McClure RR, Jones MR, Coleman TC, Breathitt L. Carotid angioplasty and stenting versus carotid endarterectomy: randomized trial in a community hospital. J Am Coll Cardiol 2001;38(6):1589–​95. 76. Naylor AR, Bolia A, Abbott RJ, et al. Randomized study of carotid angioplasty and stenting versus carotid endarterectomy: a stopped trial. J Vasc Surg 1998;28(2):326–​34.

699

70

700

Section 8  

non-coronary percutaneous interventions

77. Coward LJ, McCabe DJ, Ederle J, Featherstone RL, Clifton A, Brown MM. Long-​term outcome after angioplasty and stenting for symptomatic vertebral artery stenosis compared with medical treatment in the Carotid And Vertebral Artery Transluminal Angioplasty Study (CAVATAS): a randomized trial. Stroke 2007;38(5):1526–​30. 78. Ederle J, Bonati LH, Dobson J, et al.; CAVATAS Investigators. Endovascular treatment with angioplasty or stenting versus endarterectomy in patients with carotid artery stenosis in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS): long-​term follow-​up of a randomised trial. Lancet Neurol 2009;8(10):898–​907. 79. White R, for the CaRESS Steering Committee. Carotid Revascularization Using Endarterectomy or Stenting Systems (CaRESS) phase I clinical trial: 1-​year results. J Vasc Surg 2005;42(2):213–​19. 80. Yadav JS, Wholey MH, Kuntz RE, et al. Protected carotid-​artery stenting versus endarterectomy in high-​risk patients. N Engl J Med 2004;351(15):1493–​501. 81. Mas JL, Chatellier G, Beyssen B, et al. Endarterectomy versus stenting in patients with symptomatic severe carotid stenosis. N Engl J Med 2006;355(16):1660–​71. 82. Mas JL, Trinquart L, Leys D, et al. Endarterectomy Versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-​3S) trial: results up to 4 years from a randomised, multicentre trial. Lancet Neurol 2008;7(10):885–​92. 83. Ringleb PA, Allenberg J, Bruckmann H, et al. 30 day results from the SPACE trial of stent-​protected angioplasty versus carotid endarterectomy in symptomatic patients: a randomised non-​inferiority trial. Lancet 2006;368(9543):1239–​47. 84. Eckstein HH, Ringleb P, Allenberg JR, et al. Results of the Stent-​ Protected Angioplasty versus Carotid Endarterectomy (SPACE) study to treat symptomatic stenoses at 2 years: a multinational, prospective, randomised trial. Lancet Neurology 2008;7(10):893–​902. 85. Brott TG, Hobson RW 2nd, Howard G, et al.; CREST Investigators. Stenting versus endarterectomy for treatment of carotid-​artery stenosis. N Engl J Med 2010;363(1):11–​23. 86. Brott TG, Howard G, Roubin GS, et al.; CREST Investigators. Long-​ term results of stenting versus endarterectomy for carotid-​artery stenosis. N Engl J Med 2016;374(11):1021–​31. 87. Halliday A, Harrison M, Hayter E, et al. 10-​Year stroke prevention after successful carotid endarterectomy for asymptomatic stenosis (ACST-​ 1): a multicentre randomised trial. Lancet 2010;376:1074–​84. 88. Bonati LH, Dobson J, Featherstone RL, et al.; International Carotid Stenting Study Investigators. Long-​term outcomes after stenting versus endarterectomy for treatment of symptomatic carotid stenosis: the International Carotid Stenting Study (ICSS) randomised trial. Lancet 2015;385(9967):529–​38. 89. Doig D, Turner EL, Dobson J, et al.; ICSS Investigators predictors of stroke, myocardial infarction or death within 30 days of carotid artery stenting: results from the International Carotid Stenting Study. Eur J Vasc Endovasc Surg 2016;51(3):327–​34. 90. Gargiulo G, Sannino A, Stabile E, Perrino C, Trimarco B, Esposito G. New cerebral lesions at magnetic resonance imaging after carotid artery stenting versus endarterectomy: an updated meta-​analysis. PLoS One 2015;10(5):e0129209. 91. Kougias P, Collins R, Pastorek N, et al. Comparison of domain-​specific cognitive function after carotid endarterectomy and stenting. J Vasc Surg 2015;62(2):355–​61. 92. Lei Zhang, Zhiqing Zhao, Yaoming Ouyang, et al. Systematic review and meta-​analysis of carotid artery stenting versus endarterectomy for carotid stenosis: a chronological and worldwide study. Medicine (Baltimore) 2015;94(26):e1060. 93. Gray WA, Chaturvedi S, Verta P. Thirty-​day outcomes for carotid artery stenting in 6320 patients from two prospective, multicentre, high-​surgical-​risk registries. Circ Cardiovasc Intervent 2009;2:159–​60. 94. Rosenfield K, Matsumura JS, Chaturvedi S, et al., for the ACT Investigators. Randomized trial of stent versus surgery for asymptomatic carotid stenosis. N Engl J Med 2016;374:1011–​102.

95. Gaines PA, Randall MS. Carotid artery stenting for patients with asymptomatic carotid disease (and news on TACIT). Eur J Vasc Endovasc Surg 2005;30(5):461–​3. 96. ACST-​2 Collaborative Group, Halliday A, Bulbulia R, Gray W, et al. Status update and interim results from the asymptomatic carotid surgery trial-​2 (ACST-​2). Eur J Vasc Endovasc Surg 2013;46(5):510–​18. 97. Reiff T, Stingele R, Eckstein HH, et al. Stent-​protected angioplasty in asymptomatic carotid artery stenosis vs endarterectomy: SPACE2—​ a three-​arm randomized-​controlled clinical trial. Int J Stroke 2009;4:294–​9. 98. European Carotid Surgery Trial-​2 website. Retrieved 2016 from www. ecst-​2.com 99. CREST-​2: https://​clinicaltrials.gov/​ct2/​show/​NCT02089217 100. McCabe DJ, Pereira AC, Clifton A, Bland JM, Brown MM. Restenosis after carotid angioplasty, stenting, or endarterectomy in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS). Stroke 2005;36(2):281–​6. 101. Lal BK, Hobson RW, Goldstein J, et al. In-​stent recurrent stenosis after carotid artery stenting: life table analysis and clinical relevance. J Vasc Surg 2003;38(6):1162–​8. 102. Bosiers M, Peeters P, Deloose K, et al. Does carotid artery stenting work on the long run: 5-​year results in high-​volume centers (ELOCAS Registry). J Cardiovasc Surg (Torino) 2005;46(3):241–​7. 103. Lal BK, Kaperonis EA, Cuadra S, Kapadia I, Hobson RW. Patterns of in-​stent restenosis after carotid artery stenting: classification and implications for long-​term outcome. J Vasc Surg 2007;46(5):833–​40. 104. Lal BK. Recurrent carotid stenosis after CEA and CAS: diagnosis and management. Semin Vasc Surg 2007;20(4):259–​66. 105. Lal BK, Hobson RW, Tofighi B, Kapadia I, Cuadra S, Jamil Z. Duplex ultrasound velocity criteria for the stented carotid artery. J Vasc Surg 2008;47(1):63–​73. 106. Zhou W, Lin PH, Bush RL, et al. Management of in-​sent restenosis after carotid artery stenting in high-​risk patients. J Vasc Surg 2006;43(2):305–​12. 107. Setacci C, De Donato G, Setacci F, et al. In-​stent restenosis after carotid angioplasty and stenting: a challenge for the vascular surgeon. Eur J Vasc Endovasc Surg 2005;29(6):601–​7. 108. Akin E, Knobloch K, Pichlmaier M, Haverich A. Instent restenosis after carotid stenting necessitating open carotid surgical repair. Eur J Cardiothorac Surg 2004;26(2):442–​3. 109. Hynes BG, Kennedy KF, Ruggiero NJ, et al. Carotid artery stenting for recurrent carotid artery restenosis after previous ipsilateral carotid artery endarterectomy or stenting: a report from the National Cardiovascular Data Registry. J Am Coll Cardiol Intv 2014;7(2):180–​6. 110. Sangiorgi G, Romagnoli E, Biondi-​Zoccai G. Drug-​eluting balloons for carotid in-​stent restenosis: can this technology delivery the goods. J Endovasc Ther 2012;19:743–​8. 111. Tekieli L, Pieniazek P, Musialek P, et al. Zotarolimus-​eluting stent for the treatment of recurrent, severe carotid artery in-​stent stenosis in the TARGET-​CAS population. Endovasc Ther 2012;19(3):316–​24. 112. Giordano A, Ferraro P, Corcione N, et al. Successful treatment of recurrent carotid in-​stent restenosis and drug-​eluting balloon failure with a coronary bioresorbable vascular scaffold: a case report. Int J Surg Case Rep 2016;21:78–​82. 113. Nederkoorn PJ, van der GY, Hunink MG. Duplex ultrasound and magnetic resonance angiography compared with digital subtraction angiography in carotid artery stenosis: a systematic review. Stroke 2003;34(5):1324–​32. 114. Jahromi AS, Cina CS, Liu Y, Clase CM. Sensitivity and specificity of color duplex ultrasound measurement in the estimation of internal carotid artery stenosis: a systematic review and meta-​analysis. J Vasc Surg 2005;41(6):962–​72. 115. Biasi GM, Froio A, Diethrich EB, et al. Carotid plaque echolucency increases the risk of stroke in carotid stenting: the Imaging in

701

Chapter 46 

116. 117. 118. 119. 120. 121. 122. 123. 124. 125.

126. 127. 128. 129.

130. 131. 132. 133.

134. 135. 136.

Carotid Angioplasty and Risk of Stroke (ICAROS) study. Circulation 2004;110(6):756–​62. Reiter M, Bucek RA, Effenberger I, et al. Plaque echolucency is not associated with the risk of stroke in carotid stenting. Stroke 2006;37(9):2378–​80. Davies KN, Humphrey PR. Complications of cerebral angiography in patients with symptomatic carotid territory ischaemia screened by carotid ultrasound. J Neurol Neurosurg Psychiatry 1993;56(9):967–​72. Rothwell PM, Pendlebury ST, Wardlaw J, Warlow CP. Critical appraisal of the design and reporting of studies of imaging and measurement of carotid stenosis. Stroke 2000;31(6):1444–​50. Koelemay MJ, Nederkoorn PJ, Reitsma JB, Majoie CB. Systematic review of computed tomographic angiography for assessment of carotid artery disease. Stroke 2004;35(10):2306–​12. Saba L, Anzidei M, Piga M, et al. Multi-​modal CT scanning in the evaluation of cerebrovascular disease patients. Cardiovasc Diagn Ther 2014;4(3):245–​62. Cates CU, Patel AD, Nicholson WJ. Use of virtual reality simulation for mission rehearsal for carotid stenting. JAMA 2007;297(3):265–​6. Briguori C, Donnarumma E, Quintavalle C, Fiore D, Condorelli G. Contrast-​induced acute kidney injury: potential new strategies. Curr Opinions Nephrol Hypertens 2015;24(2):145–​53. Lin PH, Bush RL, Peden EK, et al. Carotid artery stenting with neuroprotection: assessing the learning curve and treatment outcome. Am J Surg 2005;190(6):850–​7. MacLean AA, Peta CS, Katzen BT. Bivalirudin in peripheral interventions. Tech Vasc Interv Radiol 2006;9(2):80–​3. Zahn R, Ischinger T, Hochadel M, et al. Glycoprotein IIb/​IIIa antagonists during carotid artery stenting: results from the carotid artery stenting (CAS) registry of the Arbeitsgemeinschaft Leitende Kardiologische Krankenhausarzte (ALKK). Clin Res Cardiol 2007;96(10):730–​7. Ho DS, Wang Y, Chui M, Wang Y, Ho SL, Cheung RT. Intracarotid abciximab injection to abort impending ischemic stroke during carotid angioplasty. Cerebrovasc Dis 2001;11(4):300–​4. Arab D, Yahia AM, Qureshi AI. Use of intravenous abciximab as adjunctive therapy for carotid angioplasty and stent placement. Int J Cardiovasc Intervent 2003;5(2):61–​6. Castriota F, Cremonesi A, Manetti R, Lamarra M, Noera G. Carotid stenting using radial artery access. J Endovasc Surg 1999;6(4):385–​6. Ruzsa Z, Nemes B, Pintér L, et al. A randomised comparison of transradial and transfemoral approach for carotid artery stenting: RADCAR (RADial access for CARotid artery stenting) study. EuroIntervention 2014;10(3):381–​91. Feldtman RW, Buckley CJ, Bohannon WT. How I do it: cervical access for carotid artery stenting. Am J Surg 2006;192(6):779–​81. Rafii-​Tari H, Riga CV, Payne CJ, et al. Reducing contact forces in the arch and supra-​aortic vessels using the Magellan robot. J Vasc Surg 2016;64(5):1422–​32. Ohki T, Marin ML, Lyon RT, et al. Ex vivo human carotid artery bifurcation stenting: correlation of lesion characteristics with embolic potential. J Vasc Surg 1998;27(3):463–​71. Jordan WD, Jr., Voellinger DC, Doblar DD, Plyushcheva NP, Fisher WS, McDowell HA. Microemboli detected by transcranial Doppler monitoring in patients during carotid angioplasty versus carotid endarterectomy. Cardiovasc Surg 1999;7(1):33–​8. Bosiers M, Deloose K, Verbist J, Peeters P. What practical factors guide the choice of stent and protection device during carotid angioplasty? Eur J Vasc Endovasc Surg 2008;35(6):637–​43. Wholey MH, Al Mubarek N, Wholey MH. Updated review of the global carotid artery stent registry. Catheter Cardiovasc Interv 2003;60(2):259–​66. Lo CH, Doblas M, Criado E. Advantages and indications of transcervical carotid artery stenting with carotid flow reversal. J Cardiovasc Surg (Torino) 2005;46(3):229–​39.

carotid artery stenting

137. Cano MN, Kambara AM, de Cano SJ, et al. Randomized comparison of distal and proximal cerebral protection during carotid artery stenting. JACC Cardiovasc Interv 2013;6(11):1203–​9. 138. El Koussy M, Schroth G, Do DD, et al. Periprocedural embolic events related to carotid artery stenting detected by diffusion-​weighted MRI: comparison between proximal and distal embolus protection devices. J Endovasc Ther 2007;14(3):293–​303. 139. Kwolek CJ, Jaff MR, Leal JI, et al. Results of the ROADSTER multicenter trial of transcarotid stenting with dynamic flow reversal. Vasc Surg 2015;62(5):1227–​34. 140. Bosiers M, Deloose K, Verbist J, Peeters P. The impact of embolic protection device and stent design on the outcome of CAS. Perspect Vasc Surg Endovasc Ther 2008;20(3):272–​9. 141. Hobson RW, Lal BK, Chakhtoura E, et al. Carotid artery stenting: analysis of data for 105 patients at high risk. J Vasc Surg 2003;37(6):1234–​9. 142. Jansen O, Fiehler J, Hartmann M, Bruckmann H. Protection or nonprotection in carotid stent angioplasty. The influence of interventional techniques on outcome data from the SPACE trial. Stroke 2009;40(3):841–​6. 143. Bosiers M, De Donato G, Deloose K, et al. Does free cell area influence the outcome in carotid artery stenting? Eur J Vasc Endovasc Surg 2007;33(2):135–​41. 144. Kouvelos GN, Patelis N, Antoniou GA, Lazaris A, Matsagkas MI. Meta-​analysis of the effect of stent design on 30-​day outcome after carotid artery stenting. J Endovasc Ther 2015;22(5):789–​97. 145. Schofer J, Musiałek P, Bijuklic K, et al. A prospective, multicenter study of a novel mesh-​covered carotid stent: the CGuard CARENET Trial (Carotid Embolic Protection Using MicroNet). JACC Cardiovasc Interv 2015;8(9):1229–​34. 146. Hopf-​Jensen S, Marques L, Preiß M, Müller-​Hülsbeck S. Initial clinical experience with the micromesh roadsaver carotid artery stent for the treatment of patients with symptomatic carotid artery disease. J Endovasc Ther 2015;22(2):220–​5. 147. Cieri E, De Rango P, Maccaroni MR, Spaccatini A, Caso V, Cao P. Is haemodynamic depression during carotid stenting a predictor of peri-​procedural complications? Eur J Vasc Endovasc Surg 2008;35(4):399–​404. 148. Qureshi AI, Luft AR, Sharma M, et al. Frequency and determinants of postprocedural hemodynamic instability after carotid angioplasty and stenting. Stroke 1999;30(10):2086–​93. 149. Coutts SB, Hill MD, Hu WY. Hyperperfusion syndrome: toward a stricter definition. Neurosurgery 2003;53(5):1053–​8. 150. Berjljung L, Hjorth S, Svendler CA, Oden B. Angiography in acute gastrointestinal bleeding. Surg Gynecol Obstet 1977;145(4):501–​3. 151. Ecker RD, Guidot CA, Hanel RA, et al. Perforation of external carotid artery branch arteries during endoluminal carotid revascularization procedures: consequences and management. J Invasive Cardiol 2005;17(6):292–​5. 152. Broadbent LP, Moran CJ, Cross DT, III, Derdeyn CP. Management of ruptures complicating angioplasty and stenting of supraaortic arteries: report of two cases and a review of the literature. AJNR Am J Neuroradiol 2003;24(10):2057–​61. 153. Puñal-​Riobóo J, Atienza G, Blanco M. Safety and efficacy of mechanical thrombectomy using stent retrievers in the endovascular treatment of acute ischaemic stroke: a systematic review. Interv Neurol 2015;3(3–​4):149–​64. 154. Tsang JS, Naughton PA, Leong S, Hill AD, Kelly CJ, Leahy AL. Virtual reality simulation in endovascular surgical training. Surgeon 2008;6(4):214–​20. 155. Scott DJ, Dunnington GL. The new ACS/​APDS Skills Curriculum: moving the learning curve out of the operating room. J Gastrointest Surg 2008;12(2):213–​21. 156. Van Herzeele I, Aggarwal R, Choong A, Brightwell R, Vermassen FE, Cheshire NJ. Virtual reality simulation objectively differentiates level of carotid stent experience in experienced interventionalists. J Vasc Surg 2007;46(5):855–​63.

701

702

702

Section 8  

non-coronary percutaneous interventions

157. Dawson S, Gould DA. Procedural simulation’s developing role in medicine. Lancet 2007;369(9574):1671–​3. 158. Van Herzeele I, Aggarwal R, Neequaye S, et al. Experienced endovascular interventionalists objectively improve their skills by attending carotid artery stent training courses. Eur J Vasc Endovasc Surg 2008;35(5):541–​50. 159. Katzen BT, Criado FJ, Ramee SR, et al. Carotid artery stenting with emboli protection surveillance study: thirty-​day results of the CASES-​ PMS study. Catheter Cardiovasc Interv 2007;70(2):316–​23. 160. Carotid Stenting Trialists’ Collaboration. Short-​term outcome after stenting versus endarterectomy for symptomatic carotid stenosis: a preplanned meta-​analysis of individual patient data. Lancet 2010;376:1062–​73.

161. Calvet D, Mas JL, Algra A, et al. Carotid stenting: is there an operator effect? A pooled analysis from the carotid stenting trialists’ collaboration. Stroke 2014;45:527–​32. 162. Healey P, Samanta J. When does the ‘learning curve’ of innovative interventions become questionable practice? Eur J Vasc Endovasc Surg 2008;36:253–​7. 163. MacDonald S, Lee R, Williams R, Stansby G, on behalf of the Dephi Carotid Stenting Consensus Panel. Towards safer carotid artery stenting: a scoring system for anatomic suitability. Stroke 2009;40:1698–​703.

703

CHAPTER 47

Left atrial appendage occlusion Sandeep Panikker, Tim Betts, and Milena Leo

Introduction Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, affecting 1.5–​2% of the general population and more than 8% of those older than 80 years (1, 2). Because of the progressive ageing of our population, an exponential increase in incidence is expected over the next few decades (3). Patients with AF have an increased mortality and morbidity, particularly owing to fatal or disabling stroke. The risk of embolic stroke is five times higher in the presence of AF, with an average annual rate around 5% but a progressive increase with age and the presence of other risk factors, such as prior stroke or transient ischaemic attack, hypertension, diabetes mellitus, congestive heart failure, female sex, and vascular disease, as predicted by the CHADS2 and the CHA2DS2-​VASc scores (4–​7). Moreover, strokes associated with AF are more severe, with a 50% greater likelihood of becoming disabled or handicapped and more than 50% likelihood of death (8, 9). Intracardiac thrombus formation due to the Virchow triad of events (endothelial or endocardial damage or dysfunction, abnormal blood stasis, and altered haemostasis, platelet function, and fibrinolysis), followed by distal embolization leads to thromboembolic events manifest as transient ischaemic attack (TIA), ischaemic stroke, and peripheral embolism in patients with AF (10). As the clotting cascade is integral to thrombus formation, stroke prevention in AF patients has traditionally been based on oral anticoagulant therapy with vitamin K antagonists (VKAs). Different trials and meta-​analyses conducted within the past two decades have demonstrated the efficacy of warfarin therapy in this setting, with a 65% reduction of cerebrovascular events in comparison with placebo and a 40% reduction relative to oral antiplatelet agents (11, 12). However, despite the clear advantage, this treatment is underutilized in patients with AF owing to poor patient compliance, contraindications, potential or previous bleeding complications, and interaction with food or other drugs (13, 14). Even in suitable and compliant patients, the therapeutic range is achieved and maintained only in 50–​70% of monitored days (15). Non-​vitamin K antagonist oral anticoagulants (NOACs) have been developed to overcome the VKA-​associated disadvantages and are now available for stroke prevention in patients with AF. Unfortunately, the risk of major bleeding with these new agents remains, thus increasing interest in alternative therapies for stroke prevention (16–​19).

Rationale for left atrial appendage exclusion Transoesophageal echocardiography (TOE) and pathological studies have shown that more than 90% of emboli in non-​valvular

AF originate in the left atrial appendage (LAA) (20), a remnant of the embryonic left atrium (LA), lying in the left atrioventricular groove in close relation with the left circumflex artery, the left superior pulmonary vein posteriorly, the mitral valve annulus medially, and the left phrenic nerve laterally. In contrast to the rest of the LA, which has a smooth endocardial surface, the LAA presents numerous endocardial trabeculae (pectinate muscles) and often a multilobed structure forming crypts that can harbour blood clots. Moreover, the LAA is very distensible, thus receiving a large volume of blood. These anatomical features, in combination with the significantly reduced contractility, lead to blood stagnation and thrombosis preferentially in the LAA during AF (21) (Fig 47.1). For these reasons, LAA exclusion has been developed over the years as a potential alternative to anticoagulation for stroke prevention in patients with AF (7). It is important to clarify that this paradigm applies to non-​ valvular AF, defined as AF that is not in the setting of rheumatic mitral valve disease or mitral valve surgery. In rheumatic mitral stenosis there is significant endothelial pathology, and thrombus formation frequently occurs outside of the LAA; in this instance only VKAs have been shown to be efficacious (22, 23).

Available techniques for left atrial appendage exclusion, with focus on clinical evidence Three general approaches have been developed to exclude the LAA: 1. A  surgical approach directed at amputation or ligation of the LAA. 2. A percutaneous endovascular strategy consisting of deployment of an occlusion device inside the LAA. 3. A percutaneous epicardial ligation technique aimed at externally excluding the LAA (Fig 47.2).

The surgical approach The first report of LAA resection during mitral valvulotomy dates back to 1949 (24). Since then, different techniques have been developed to exclude the LAA: ligation by endocardial and/​or epicardial suturing or stapling, stapling excision, or surgical excision followed by oversew (25). At the beginning the popularity of the procedure was low, owing to a high failure rate (high incidence of residual flow and residual LAA stump) and to the uncertain clinical relevance because a large

704

704

Section 8  

non-coronary percutaneous interventions

LA LAA

LA

LV

LAA

AV

A

B

C

D

LAA

Figure 47.1  Left atrial appendage (LAA) imaged by transoesophageal echocardiography; patients in atrial fibrillation. A) Severely dilated left atrium, with spontaneous echo-​contrast as per blood stasis. B) Multilobed structure of the LAA (black arrows). C) Endocardial trabeculae (pectinate muscles) in the LAA. D) Reduced LAA emptying velocity during atrial fibrillation. AV, Aortic valve; LA, left atrium; LAA, left atrial appendage; LV, left ventricle.

LEFT ATRIAL APPENDAGE EXCLUSION TECHNIQUES SURGICAL APPROACH

PERCUTANEOUS APPROACH

EXCISION

ENDO/EPICARDIAL LIGATION 1. Surgical cut and oversew

LARIAT device

2. Stapling LIGATION

ENDOCARDIAL OCCLUSION DEVICES

1. Endo/epicardial suturing 2. Stapling

WATCHMAN

ACP

Figure 47.2  Available techniques and devices for left atrial appendage exclusion.

WAVECREST

AMULET

705

Chapter 47 

number of patients needed lifelong anticoagulation therapy for indications unrelated to AF. As documented by the LAAOS I (26) and LAAOS II trials (27), the surgical techniques for LAA exclusion have progressively improved over time, with increased closure success rates. The AtriClip (Atricure, Inc, West Chester, OH, USA) has been recently developed to exclude the LAA during open cardiac surgery. Preliminary data from small studies have showed that the AtriClip implant is feasible and safe, with a high rate of successful implants, no procedural adverse events, and complete LAA closure in up to 98% of cases (28). Despite improvements in the technologies, data on long-​term stroke prevention by surgical LAA exclusion are scarce and controversial. A large randomized trial (LAAOS III) is currently recruiting participants to evaluate the safety and efficacy of LAA removal in patients with AF undergoing concomitant heart surgery (29).

The percutaneous endovascular approach Exclusion of the LAA cavity from the atrium can be achieved with occlusion devices that are delivered percutaneously through transseptal access to the LA. To date, five devices have been investigated for percutaneous LAA occlusion (LAAO); four of them are currently available for clinical use and have a CE mark. The PLAATO (Percutaneous Left Atrial Appendage Transcatheter Occlusion; Ev3, Plymouth, MN, USA) was the first LAAO device to be designed (30). It consisted of a self-​expanding nitinol cage, with an impermeable polymeric membrane covering the surface facing the LA and three rows of anchors helping to stabilize the device in the LAA (Fig 47.3A). Despite favourable clinical results, it has been discontinued for commercial reasons. Morphologically similar to the PLAATO, the Watchman (Boston Scientific, Maple Grove, MN, USA) consists of a self-​expanding parachute-​shaped nitinol cage, with a row of fixation barbs around the surface and a permeable polyethylene terephthalate (PET) membrane on the surface facing the LA (Fig 47.3B). Of note, the Watchman is the only LAAO device that has been evaluated in prospective, randomized controlled trials (RCTs) and has Food and Drug Administration (FDA) approval. The PROTECT AF trial (31, 32)  was the first and largest RCT to evaluate the non-​inferiority of an LAAO therapy for stroke risk reduction. It enrolled 707 patients with non-​valvular AF and eligible to warfarin who were randomized to Watchman or warfarin in a 2:1 ratio. Patients in the device group received warfarin for 15 days before and a minimum of 45 days following implantation. TOE was performed at 45  days, 6  months, and 12  months to assess any residual peridevice flow. Warfarin was discontinued if the LAA closure was complete or the width of the flow jet was 7 days after randomization. On these premises, the Watchman device has recently received FDA approval for patients at increased risk of stroke and systemic embolism who are suitable for short-​term warfarin therapy but have an appropriate rationale to seek a long-​term non-​pharmacological alternative to warfarin. Although limited data are available about the use of the Watchman device in patients with a contraindication to anticoagulation, the previous experience with the PLAATO device (30) and registries such as the ASAP study (37) and EWOLUTION (38) suggest that it is feasible, particularly as these are the patients in whom the device is currently recommended in national and international guidelines. The ASAP (ASA Plavix Feasibility Study with Watchman Left Atrial Appendage Closure Technology) study (37) evaluated 150 patients with non-​valvular AF and CHADS2 score ≥1 who were ineligible for warfarin. Patients were treated with aspirin and clopidogrel for 6  months, and then lifelong aspirin after Watchman implantation. The mean follow-​up duration was 14 months. The incidence of stroke and systemic embolism was 2.3% per year, 77% lower if compared with the expected ischaemic stroke rate based on the CHADS2 score. The EWOLUTION registry (38) included all subjects elected to a Watchman implant, according to the appropriate local and international guidelines, in 47 European participating centres (with varying levels of past experience) between October 2013 and May 2015. When compared with previous studies, the 1025 subjects had a higher stroke risk and bleeding risk (average CHADS2 score 2.8, average CHA2DS2-​VASc score 4.5, HAS-​BLED score ≥3 in 40% of subjects), and 62% were deemed unsuitable for anticoagulation. Device deployment was successful in 98.5% of cases, with a very low incidence of periprocedural complications (2.8% at 7  days post-​procedure, 3.6% at 30 days). The Amplatzer Cardiac Plug (ACP-​ 1; St. Jude Medical, Minneapolis, MN, USA) is a nitinol self-​expanding device consisting of two parts, a lobe and a disc which are connected by a central waist. Twelve stabilizing wires are equally displaced around the main disc and contribute to device retention and stabilization inside the LAA (Fig 47.3C). The lobe enters into the LAA neck

705

706

706

Section 8  

non-coronary percutaneous interventions

A

B

C

D

E

F

Figure 47.3  Different percutaneous left atrial appendage exclusion devices. A) PLAATO; B) Watchman; C) Amplatzer Cardiac Plug; D) Amplatzer Amulet; E) WaveCrest; F) LARIAT.

with a depth of 10 mm or more, anchoring the device, whereas the disc covers the LAA orifice, thereby occluding it. Anticoagulation therapy is not required after the device implant:  dual antiplatelet therapy is recommended for 1–​6 months (39–​41), followed by aspirin only (42). Although the ACP is the device with the longest clinical follow-​ up among the currently available LAAO devices, no RCTs are available to allow a comparison with anticoagulation therapy. The largest registry (39) has reported a successful device implantation in 97% of cases, with a very low percentage of significant residual peridevice flow (1.9% at 7  months), procedural complications in about 5% of cases, and significant long-​term benefits in terms of both stroke risk reduction and major bleeding risk reduction (59% and 61%, respectively), if compared with the predicted risks on the basis of CHA2DS2-​VASc and HAS-​BLED scores. The newest Amplatzer Amulet (ACP-​ 2; St. Jude Medical, Minneapolis, MN, USA) is the next-​generation device. The general

design, with a lobe and a disc, has been maintained but some improvements have been made to increase the deliverability and the stability of the device:  increased number and stiffer stabilizing wires, increased length of the distal lobe and diameter of the proximal disc, and availability of larger sizes (Fig 47.3D). Only small case series have been published to date, showing a successful implant in 96% of cases with no procedural complications and no strokes at 3 months (43, 44). The WaveCrest (Coherex Medical, Salt Lake City, UT, USA) is an umbrella-​shaped device consisting of a nitinol structure without exposed metal hub, covered by a foam layer on the inner surface facing the LAA and by a polymeric membrane on the surface facing the LA (Fig 47.3E). The anchors are rolled out after proximal positioning of the device, thereby allowing a very controlled release. The landing zone is generally more proximal than other devices, thus allowing treatment of proximal LAA lobe bifurcations and very angulated LAAs. Other unique features are the possibility of

70

Chapter 47 

left atrial appendage occlusion

Table 47.1  Possible indications for left atrial appendage exclusion Indication

Potential candidates

Lifestyle choice

Patients with AF and CHADS2-​VASc score ≥1

Theoretical environmental risk for anticoagulation

Competitive contact sports that could lead to a head injury (football, rugby, cricket, hockey, martial arts, kickboxing). Professions associated with a higher risk of trauma and bleeding (mason, mechanic, fisherman, policeman, farmer, etc.)

Relative contraindication to anticoagulation

High bleeding risk HAS-​BLED ≥3. Need for prolonged triple antithrombotic therapy, severe renal failure, thrombocytopenia, myelodysplastic syndrome, cancer, uncontrolled hypertension, vascular malformations associated with high bleeding risk, recurrent minor bleeding on anticoagulation (i.e. nose bleed), intolerance or reduced compliance to anticoagulant drugs

Absolute contraindication to anticoagulation

Previous life-​threatening bleeding on/​off anticoagulation (i.e. intracranial haemorrhage), recurrent gastrointestinal bleeding despite endoscopic therapy

Additional stroke risk protection

Previous thromboembolic event(s) despite well controlled anticoagulation

operating the occluder and the anchoring system independently, thus allowing repositioning before anchoring, and the possibility of distal injection (on the appendage side of the occluder) to assess stability and occlusion. It has been tested in animals and humans with satisfactory and promising results (implant success rate of 96%, 45-​day primary efficacy 97%) (45, 46).

The percutaneous epicardial approach This technique is a hybrid of the surgical and percutaneous endocardial approaches and utilizes a suture delivery device called LARIAT (SentreHEART, Inc., Palo Alto, CA, USA) that consists of a pre-​tied suture contained on a closure snare, which is guided via a catheter epicardially over the LAA (Fig 47.3F). A combined pericardial and endocardial access to the LAA is obtained through a percutaneous subxiphoid epicardial puncture and an interatrial transseptal puncture, respectively. A  magnet-​tipped wire is advanced to the LAA and a 15-​mm balloon-​tipped EndoCATH (SentreHEART, Inc.) catheter is positioned to the LAA ostium. A second magnet-​tipped wire is then advanced to the LAA epicardially and, after establishing magnetic contact with the endocardial wire, it is used to advance the LARIAT loop to the LAA. The endocardial balloon is inflated and used to help deployment of the LARIAT to the LAA neck. After having confirmed the complete LAA closure by TOE and angiography, the endocardial hardware is removed from the LAA, the suture is deployed, the LARIAT delivery device is removed from the pericardium and the suture is cut. A final angiogram is performed to document LAA closure (47). The lack of need for initial oral anticoagulation therapy makes the LARIAT technique appealing for a population of patients with previous major bleeding or with high bleeding risk. The technique is feasible, with a documented procedural success around 93–​94%, but not particularly safe: a high incidence of pericardial effusions requiring pericardiocentesis (between 11 and 20%), major bleeding (9%), and LAA perforation requiring open-​chest surgery (9%) have been documented (48, 49). Moreover, some LAA anatomical variants, such as large LAAs, posteriorly rotated LAAs, superiorly orientated LAA lobes, and posteriorly rotated hearts, can be challenging to ligate with the LARIAT device; previous cardiac surgery can preclude the use of the LARIAT because of postoperative pericardial adhesions as well. Finally, data about long-​term stroke prevention are lacking to date.

Step-​by-​step left atrial appendage exclusion: tips and tricks for a successful procedure Before the procedure Choose the right patient Whether LAA exclusion is a niche procedure for patients with absolute contraindications to oral anticoagulants (OACs) or a viable option for all who need thromboprophylaxis is still subject to debate. According to the 2012 European Society of Cardiology (ESC) guidelines on management of AF (7), percutaneous LAA closure may be considered in patients with a high stroke risk (CHADS2 score >1 and/​or CHA2DS2-​VASc score >2) and contraindications for long-​term oral anticoagulation, while surgical excision of the LAA may be considered in patients undergoing open-​heart surgery (Class IIb recommendations, Level of evidence B and C, respectively). Given the absence of controlled clinical data, this restrictive clinical indication comes from expert consensus only. More recently, the European Heart Rhythm Association/​European Association of Percutaneous Cardiovascular Interventions (EHRA/​EAPCI) consensus statement on catheter-​based LAAO (50) has suggested percutaneous LAAO in a broader spectrum of circumstances. Growing evidence suggests that LAAO is at least as effective as warfarin at reducing stroke in selected patients and lowers mortality and bleeding risk associated with lifelong anticoagulation (51). The procedure carries some risks but they decrease with operator experience (34). Of note, LAAO closure is likely to be cost-​ effective (52), particularly in younger patients with high risk of thromboembolic stroke. On the basis of these considerations, it could be reasonable to consider LAAO not only in patients with a relative or absolute contraindication to oral anticoagulation but also in the presence of theoretical environmental risks or as lifestyle choice in patients with CHA2DS2-​VASc score ≥1(53) (Table 47.1).

Collect the right information Careful estimates of risks for stroke (using the CHADS2 or the CHA2DS2-​VASc score), bleeding (using the HAS-​BLED score but also considering the presence of other bleeding predisposing conditions not addressed in the HAS-​BLED score), and of the risks and benefits of LAA exclusion compared with pharmacological

707

708

708

Section 8  

non-coronary percutaneous interventions

alternatives or no therapy should be performed to confirm the indication to the procedure. The patient should be involved in the discussion (50, 54). Once an appropriate indication for the procedure is confirmed, important clinical information should be collected when planning the procedure: ◆ Previous

experience with antiplatelet and anticoagulant therapy, in particular the specific drugs used in the past and any side effects or contraindications.

◆ Characterization

of cardiac structure and function, with particular attention to conditions requiring heart surgery.

◆ Characterization

of LAA anatomy and exclusion of LAA thrombi before the procedure. The LAA usually has a well defined oval-​ shaped orifice that leads to a neck region that opens to a main body. Different morphological types of LAA have recently been identified, with the chicken wing being the most common (48%), followed by cactus (30%), windsock (19%), and cauliflower (3%) (55). The LAA shape, the maximal diameter of the orifice (measured from the level of the circumflex coronary artery to a point 1–​ 2 cm from the tip of the left superior pulmonary vein limbus), the diameter of the landing zone (rim of tissue beyond the neck where the disc is anchored), and the maximal depth of the appendage (distance between the ostium and the apex) are important to guide the decision about the technique to choose, the particular device to use, and its size. Cardiac magnetic resonance imaging (MRI), computed tomography (CT), and TOE can be used for pre-​procedural assessment of the LAA. Among them, TOE looks to be the most advantageous because it is widely available, it has the highest sensitivity and specificity for detection of LAA thrombi, it can be used in real time during the procedure, and it does not require ionizing radiation or contrast. Of note, 3D TOE allows better delineation of the appendage and the surrounding structures (56).

Choose the right technique and the right device Although surgical ligation or amputation of the LAA is now the standard of care in patients undergoing mitral valve surgery or as an adjunct to the maze procedure (7), its widespread utilization is limited by its invasive nature and controversial efficacy. The vast majority of patients have non-​valvular AF and have no indication for cardiac surgery, making it difficult to justify an open-​chest surgical procedure for LAA exclusion. Even though small case series have recently reported the feasibility of thoracoscopic LAA ligation using an endoloop snare or stapling, percutaneous approaches should be preferred in first instance in patients not undergoing cardiac surgery for other reasons (57–​59). No randomized comparative studies are available to compare the different percutaneous devices. Registries and systematic reviews suggest comparable technical success and safety of the two leading devices (Watchman and ACP) in experienced hands, with equivalent long-​term efficacy (60–​62). Less evidence is available about the WaveCrest (45, 46)  and the Amulet devices (43, 44), but preliminary data suggest that they are similarly feasible and effective. The patient’s medical history should also be taken into account in relation to the need of even short-​term antithrombotic therapy after the device implant, with which associated bleeding risk is not negligible. No antithrombotic therapy is required after a LARIAT device implant. Fortunately, the rate of severe bleeding on short-​ term antiplatelet therapy, even in high-​risk patients, appears to be

acceptable, ranging between 1 and 2.85% (39–​41, 63) in different studies, and is lower when compared with OAC (64). Apart from a few selected cases where LAA anatomical considerations or clinical features of the patient orientate towards a particular device, most devices can be used in most patients. Therefore the choice should be primarily dictated by the experience and the familiarity of the operator with the technique, factors that are important determinants of procedural success and complications.

Choose the right team A multidisciplinary team is the key of a successful percutaneous LAAO. The initial evaluation should be performed by both a specialist able to characterize the specific risks and benefits of medical therapy and a procedural specialist able to estimate risks and benefits of the proposed procedure. A specialist in LAA imaging is also essential to confirm the absence of LAA thrombi prior to LAAO, to select the most adequate occlusion device, and to guide and confirm its correct placement. An anaesthetist should be involved in pre-​procedural evaluation, intraprocedural management, and post-​procedural management when general anaesthesia is planned. A cardiac surgeon should be available for surgical back-​up in case of emergency (50). Case selection should involve a neurologist and/​ or stroke physician.

Undertaking a percutaneous left atrial appendage occlusion procedure Choose the right equipment ◆ Intraprocedural

echocardiography. Apart from X-​ray guidance, TOE is used in most centres to rule out the presence of LAA thrombus, to facilitate the transseptal puncture, to measure the LAA dimensions to choose the right device size, to guide the device positioning and deployment, and to rule out the presence of pericardial effusion or device embolization after the implant. Intracardiac echocardiography (ICE) is an alternative, but limited experience has been reported to date (65). 3D TOE has become important to guide device delivery into the LAA (Fig 47.4) (56, 66).

◆ Anaesthetic

equipment. Although deep sedation may not be needed if the procedure time is not very long, most centres perform the procedure under general anaesthesia to avoid any patient movements and to be able to use continuous TOE.

◆ Transseptal

kit. Any transseptal kit can be used; the choice should be primarily dictated by the operator’s experience.

◆ Stiff

guide wire. This is used after the transseptal puncture to exchange the transseptal sheath with the delivery sheath of the device.

◆ Pigtail

catheter, for LAA contrast injections and sheath manipulation.

◆ Device delivery kit (sheath, catheter, and selection of device sizes).

Follow the right sequence ◆ Premedicate

the patient with antibiotics and antiplatelet loading doses (if the patient is not on warfarin).

◆ Anaesthetize ◆ Position

the patient.

the TOE or the ICE probe.

709

Chapter 47 

left atrial appendage occlusion

Left Superior Pulmonary Vein ostium

WATCHMAN in LAA

A

AMPLATZER CARDIAC PLUG in LAA

Mitral valve orifice

B Figure 47.4  Real-​time 3D transoesophageal en-​face views of occlusion devices in the left atrial appendage (LAA). A) Watchman and B) Amplatzer Cardiac Plug. ◆ Obtain

an arterial line for monitoring of systemic blood pressure and a single femoral venous access.

◆ Perform

the transseptal puncture; an inferoposterior location is preferable.

◆ Administer

intravenous heparin with a target activated clotting time (ACT) of >300 s.

◆ Exchange

the transseptal sheath for the LAA access with the delivery sheath over a stiff guide wire.

◆ Manipulate ◆

the sheath into the LAA over the pigtail catheter.

Perform LAA contrast injections in multiple X-​ray views through the pigtail catheter.

◆ Measure

the LAA, primarily with TOE but also fluoroscopy in different angiographic and echocardiographic views to choose the right device size.

◆

Prepare the device, by generous flushing within the delivery catheter to eliminate any air.

◆ Advance

the LAA.

the device through the delivery sheath and deploy into

◆ Check

for appropriate positioning and compression of the device and adequate LAA sealing via echocardiography and angiography. Check stability with a tug test.

◆ If

suboptimal, recapture then redeploy the device.

◆ When

satisfactory, release the device.

709

710

710

Section 8  

non-coronary percutaneous interventions

The critical steps for the implant of an endovascular occlusion device are summarized in Fig 47.5.

After the procedure Be aware and treat the procedural complications (if any) Possible complications associated with an LAAO procedure are:  access site complications, pericardial effusions, periprocedural strokes (mainly ischaemic owing to air embolism), and device embolization requiring retrieval. A  recent meta-​analysis of published studies, consisting mainly of early case series and trials with different devices, has documented vascular access site complications in 8.6%, followed by pericardial effusions in 4.3% of cases (with tamponade in 2.2% of cases), device embolization in 3.9% of cases, need for surgery in 2.3% of cases, and periprocedural mortality in 1.2% of cases (51). Higher periprocedural complications have been described by the 2015 EHRA survey on LAAO procedure in the 33 responding centres (67). The occurrence of ischaemic or haemorrhagic stroke varied

from 1 to 25% in two centres (10%), pericardial effusion with tamponade was reported in up to 6% of cases, and device embolization in up to 20% of cases. However, the majority of centres had implanted a limited number of devices in 2014 (no more than 30 procedures performed in 80% of centres), thus it could be speculated that, owing to the limited experience, the learning curve at some centres has not yet reached its plateau. As suggested by the CAP registry (34), increased operator experience translates into lower complication rates. Of note, the low incidence of periprocedural complications showed very recently by the EWOLUTION registry (38) (2.8% at 7 days, 3.6% at 30 days post-​procedure) confirms that adverse events can be effectively lowered with adequate attention and training.

Choose the right post-​procedural management The correct post-​ procedural management after LAAO is still debated. The PROTECT-​ AF protocol (31, 44)  suggests anticoagulation with warfarin for 6 weeks after a Watchman implant to avoid thrombus formation on the device until complete

A

B

C

D

E

F

G

H

I

Figure 47.5  Main steps of a percutaneous left atrial appendage occlusion procedure (Watchman implantation). Intraprocedural transoesophageal echocardiogram to rule out left atrial appendage (LAA) thrombosis (A), measure LAA depth and width at the ostium in different views (B), and guide the transseptal puncture (C). LAA cannulation by using pigtail catheter (D), device deployment (E), contrast injection via device delivery sheath to check for LAA occlusion (F) (fluoroscopy images). Transoesophageal echocardiogram to check the position of the device in the LAA (G), measure its size in the LAA, confirm an adequate compression (H), and check for LAA seal (I).

71

Chapter 47 

endothelialization (Fig 47.6). TOE imaging is advised at 45  days, 6 months, and 12 months to assess for device stability, peridevice leaks, and device-​related thrombus. When the 45-​day TOE reveals minimal residual peridevice flow (jet width ≤5 mm) and no device-​ related thrombus, warfarin can be stopped and replaced by dual antiplatelet therapy until the 6-​month visit, after which only aspirin is continued. In case of inadequate seal or thrombus, the patient should continue taking warfarin until an adequate seal is reached or thrombus resolves before transitioning to aspirin. Although proven to be quite safe and effective, this approach is not feasible for most patients who have a contraindication to oral anticoagulation. On the basis of the previous experience with the ASAP study (37), the current approach with the Watchman in most centres is a short period of dual antiplatelet therapy (from 6 weeks to 6 months according to centre preference) after a Watchman implant, followed by single antiplatelet therapy. A similar approach is generally adopted after an ACP implant (39–​41, 63). The need for repeat TOEs after an Amplatzer device implant is debated. In Tzikas’s registry (39), TOE was performed in 63% of patients after an average of 7 months from the procedure: a significant residual leak was observed in 1.9% of cases and this was managed without reintroduction of OAC; thrombus was observed in 4.4% of patients and this was treated with a brief period of OAC or low molecular weight heparin until resolution, dual antiplatelet therapy for life, or no treatment due to very high bleeding risk, according to the centre preference. No correlation between adverse events and residual leak or thrombus was observed. How long single antiplatelet therapy should be taken after Watchman or ACP implant is also debated. An individualized approach is adopted in most centres on the basis of the bleeding risk of the patient and the presence of concomitant diseases, as coronary artery disease represents an indication for lifelong aspirin therapy. Very few data are available about post-​procedural management after WaveCrest or LARIAT implant. Although no antithrombotic

Figure 47.6  Gross pathology in a dog of a left atrial appendage (LAA) successfully sealed using a percutaneous Watchman device. Note the complete LAA ostial coverage with glistening endocardium. Animal models suggest such coverage is complete by 60 days after implantation. Reproduced from David R. Holmes, Jr, and Robert S. Schwartz ‘Left atrial appendage occlusion eliminates the need for warfarin.’ Circulation. 2009;120:1919–​26, with permission from the American Heart Association.

left atrial appendage occlusion

therapy is theoretically required after a LARIAT implant, 55% of patients were still on warfarin at 1  year follow-​up, primarily because of residual leaks or incomplete closure (49).

Conclusions On the basis of the evidence supporting the role of the LAA as a source of thromboembolic stroke in more than 90% of cases in patients with AF, LAA exclusion has recently emerged as an alternative for stroke prevention in patients with AF and increased stroke risk (CHA2DS2-​VASc score ≥1). Effective means are now available to close or occlude the LAA. Surgical exclusion is feasible during concomitant open-​chest procedures, while percutaneous exclusion is preferred for stand-​alone cases. Percutaneous LAAO is a feasible technique; there are procedure-​related risks but they decrease with experience. Importantly, this technique (with the Watchman) has been shown to be at least as effective as warfarin at reducing stroke in selected patients and it lowers mortality. Moreover, LAAO is likely to be cost-​effective, particularly in younger patients with high risk of thromboembolic stroke. More studies are needed to document the equivalence of the different devices in terms of stroke risk prevention, to confirm the long-​term benefits and the cost-​effectiveness of this strategy for stroke prevention, to identify the potential candidates for the procedure, and the best antithrombotic management after LAAO.

References 1. Naccarelli GV, et al. Increasing prevalence of atrial fibrillation and flutter in the United States. Am J Cardiol 2009;104(11):1534–​9. 2. Camm AJ, et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association. Eur Heart J 2012;33(21):2719–​47. 3. Go AS, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA 2001;285(18):2370–​5. 4. Lin HJ, et al. Stroke severity in atrial fibrillation. The Framingham Study. Stroke 1996;27(10):1760–​4. 5. Gage BF, et al. Selecting patients with atrial fibrillation for anticoagulation: stroke risk stratification in patients taking aspirin. Circulation 2004;110(16):2287–​92. 6. Gage BF, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285(22):2864–​70. 7. Camm AJ, et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation—​developed with the special contribution of the European Heart Rhythm Association. Europace 2012;14(10):1385–​413. 8. Lamassa M, et al. Characteristics, outcome, and care of stroke associated with atrial fibrillation in Europe: data from a multicenter multinational hospital-​based registry (The European Community Stroke Project). Stroke 2001;32(2):392–​8. 9. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22(8):983–​8. 10. Watson T, Shantsila E, Lip GY. Mechanisms of thrombogenesis in atrial fibrillation: Virchow’s triad revisited. Lancet 2009;373(9658):155–​66. 11. Hart RG, Pearce LA, Aguilar MI. Meta-​analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007;146(12):857–​67.

711

712

712

Section 8  

non-coronary percutaneous interventions

12. Ezekowitz MD, et al. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation. Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation Investigators. N Engl J Med 1992;327(20):1406–​12. 13. Hylek EM, et al. Major hemorrhage and tolerability of warfarin in the first year of therapy among elderly patients with atrial fibrillation. Circulation 2007;115(21):2689–​96. 14. Nieuwlaat R, et al. Atrial fibrillation management: a prospective survey in ESC member countries: the Euro Heart Survey on Atrial Fibrillation. Eur Heart J 2005;26(22):2422–​34. 15. Connolly SJ, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361(12):1139–​51. 16. Connolly SJ, et al. Apixaban in patients with atrial fibrillation. N Engl J Med 2011;364(9):806–​17. 17. Viles-​Gonzalez JF, Fuster V, Halperin JL. New anticoagulants for prevention of stroke in patients with atrial fibrillation. J Cardiovasc Electrophysiol 2011;22(8):948–​55. 18. Gallego P, Roldan V, Lip GY. Novel oral anticoagulants in cardiovascular disease. J Cardiovasc Pharmacol Ther 2014;19(1):34–​44. 19. Patel MR, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011;365(10):883–​91. 20. Blackshear JL, Odell JA. Appendage obliteration to reduce stroke in cardiac surgical patients with atrial fibrillation. Ann Thorac Surg 1996;61(2):755–​9. 21. Qamruddin S, et al. Left atrial appendage: structure, function, imaging modalities and therapeutic options. Expert Rev Cardiovasc Ther 2010;8(1):65–​75. 22. Kaymaz C, et al. Location, size and morphological characteristics of left atrial thrombi as assessed by echocardiography in patients with rheumatic mitral valve disease. Eur J Echocardiogr 2001;2(4):270–​6. 23. Vahanian A, et al. Guidelines on the management of valvular heart disease (version 2012). Eur Heart J 2012;33(19):2451–​96. 24. Madden JL. Resection of the left auricular appendix; a prophylaxis for recurrent arterial emboli. J Am Med Assoc 1949;140(9):769–​72. 25. Aryana A, Saad EB, d’Avila A. Left atrial appendage occlusion and ligation devices: what is available, how to implement them, and how to manage and avoid complications. Curr Treat Options Cardiovasc Med 2012;14(5):503–​19. 26. Healey JS, et al. Left Atrial Appendage Occlusion Study (LAAOS): results of a randomized controlled pilot study of left atrial appendage occlusion during coronary bypass surgery in patients at risk for stroke. Am Heart J 2005;150(2):288–​93. 27. Whitlock RP, et al. Left Atrial Appendage Occlusion Study II (LAAOS II). Can J Cardiol 2013;29(11):1443–​7. 28. Ailawadi G, et al. Exclusion of the left atrial appendage with a novel device: early results of a multicenter trial. J Thorac Cardiovasc Surg 2011;142(5):1002–​9, 1009 e1. 29. Whitlock R, et al. Rationale and design of the Left Atrial Appendage Occlusion Study (LAAOS) III. Ann Cardiothorac Surg 2014;3(1): 45–​54. 30. Bayard YL, et al., PLAATO (Percutaneous Left Atrial Appendage Transcatheter Occlusion) for prevention of cardioembolic stroke in non-​anticoagulation eligible atrial fibrillation patients: results from the European PLAATO study. EuroIntervention 2010;6(2):220–​6. 31. Fountain RB, et al. The PROTECT AF (WATCHMAN Left Atrial Appendage System for Embolic PROTECTion in Patients with Atrial Fibrillation) trial. Am Heart J 2006;151(5):956–​61. 32. Holmes DR, et al. Percutaneous closure of the left atrial appendage versus warfarin therapy for prevention of stroke in patients with atrial fibrillation: a randomised non-​inferiority trial. Lancet 2009;374(9689):534–​42. 33. Viles-​Gonzalez JF, et al. The clinical impact of incomplete left atrial appendage closure with the Watchman Device in patients with atrial fibrillation: a PROTECT AF (Percutaneous Closure of the Left Atrial Appendage Versus Warfarin Therapy for Prevention of Stroke in Patients With Atrial Fibrillation) substudy. J Am Coll Cardiol 2012;59(10):923–​9.

34. Reddy VY, et al. Safety of percutaneous left atrial appendage closure: results from the Watchman Left Atrial Appendage System for Embolic Protection in Patients with AF (PROTECT AF) clinical trial and the Continued Access Registry. Circulation 2011;123(4):417–​24. 35. Reddy VY, et al. Percutaneous left atrial appendage closure vs warfarin for atrial fibrillation: a randomized clinical trial. JAMA 2014;312(19):1988–​98. 36. Holmes DR, Jr., et al. Prospective randomized evaluation of the Watchman Left Atrial Appendage Closure device in patients with atrial fibrillation versus long-​term warfarin therapy: the PREVAIL trial. J Am Coll Cardiol 2014;64(1):1–​12. 37. Reddy VY, et al. Left atrial appendage closure with the Watchman device in patients with a contraindication for oral anticoagulation: the ASAP study (ASA Plavix Feasibility Study With Watchman Left Atrial Appendage Closure Technology). J Am Coll Cardiol 2013;61(25):2551–​6. 38. Boersma LV, et al. Implant success and safety of left atrial appendage closure with the WATCHMAN device: peri-​procedural outcomes from the EWOLUTION registry. Eur Heart J 2016;37:2465–​74. 39. Tzikas A, et al. Left atrial appendage occlusion for stroke prevention in atrial fibrillation: multicentre experience with the AMPLATZER Cardiac Plug. EuroIntervention 2015;11(10):1170–​9. 40. Kebernik J. et al. Safety and efficacy of left atrial appendage closure with the amplatzer cardiac plug in very high stroke and bleeding risk patients with non-​valvular atrial fibrillation. Cardiol Ther 2015;4(2):167–​77. 41. Urena M, et al. Percutaneous left atrial appendage closure with the AMPLATZER cardiac plug device in patients with nonvalvular atrial fibrillation and contraindications to anticoagulation therapy. J Am Coll Cardiol 2013;62(2):96–​102. 42. Berti S, et al. Tools and techniques clinical: transcatheter closure of left atrial appendage using the Amplatzer Cardiac Plug. EuroIntervention 2013;9(4):524–​6. 43. Lam SC, et al. Left atrial appendage closure using the Amulet device: an initial experience with the second generation amplatzer cardiac plug. Catheter Cardiovasc Interv 2015;85(2):297–​303. 44. Freixa X, et al. Left atrial appendage occlusion: initial experience with the Amplatzer Amulet. Int J Cardiol 2014;174(3):492–​6. 45. Reddy V. Clinical Experience with Coherex WaveCrest Left Atrial Appendage Occlusion system. Washington, DC, TCT, September 16, 2014. 46. Cheng Y, et al. TCT-​764 safety and biocompatibility of the coherex WaveCrest™ left atrial appendage occluder in a 30-​day canine study. J Am Coll Cardiol 2012;60:17S. 47. Srivastava MC, et al. A review of the LARIAT device: insights from the cumulative clinical experience. Springerplus 2015;4:522. 48. Bartus K, et al. Percutaneous left atrial appendage suture ligation using the LARIAT device in patients with atrial fibrillation: initial clinical experience. J Am Coll Cardiol 2013;62(2):108–​18. 49. Price MJ, et al. Early safety and efficacy of percutaneous left atrial appendage suture ligation: results from the U.S. transcatheter LAA ligation consortium. J Am Coll Cardiol 2014;64(6):565–​72. 50. Meier B, et al. EHRA/​EAPCI expert consensus statement on catheter-​ based left atrial appendage occlusion. Europace 2014;16(10):1397–​416. 51. Bajaj NS, et al. Percutaneous left atrial appendage occlusion for stroke prophylaxis in nonvalvular atrial fibrillation: a systematic review and analysis of observational studies. JACC Cardiovasc Interv 2014;7(3):296–​304. 52. Singh SM, Micieli A, Wijeysundera HC. Economic evaluation of percutaneous left atrial appendage occlusion, dabigatran, and warfarin for stroke prevention in patients with nonvalvular atrial fibrillation. Circulation 2013;127(24):2414–​23. 53. Alli O, Asirvatham S, Holmes DR, Jr. Strategies to incorporate left atrial appendage occlusion into clinical practice. J Am Coll Cardiol 2015;65(21):2337–​44. 54. Masoudi FA, et al. 2015 ACC/​HRS/​SCAI left atrial appendage occlusion device societal overview. Heart Rhythm 2015;12(10):e122–​36.

713

Chapter 47 

55. Di Biase L, et al. Does the left atrial appendage morphology correlate with the risk of stroke in patients with atrial fibrillation? Results from a multicenter study. J Am Coll Cardiol 2012;60(6):531–​8. 56. Beigel R, et al. The left atrial appendage: anatomy, function, and noninvasive evaluation. JACC Cardiovasc Imaging 2014;7(12):1251–​65. 57. Ohtsuka T, et al. Thoracoscopic stand-​alone left atrial appendectomy for thromboembolism prevention in nonvalvular atrial fibrillation. J Am Coll Cardiol 2013;62(2):103–​7. 58. Guerra M, Martins D, Miranda J. Thoracoscopic left atrial appendectomy. Rev Port Cir Cardiothorac Vasc 2013;20(4):199–​201. 59. Blackshear JL, et al. Thoracoscopic extracardiac obliteration of the left atrial appendage for stroke risk reduction in atrial fibrillation. J Am Coll Cardiol 2003;42(7):1249–​52. 60. Romero J, et al. Left atrial appendage isolation using percutaneous (endocardial/​epicardial) devices: pre-​clinical and clinical experience. Trends Cardiovasc Med 2016;26:182–​99. 61. Chun KR, et al. Left atrial appendage closure followed by 6 weeks of antithrombotic therapy: a prospective single-​center experience. Heart Rhythm 2013;10(12):1792–​9. 62. Mazzone C, et al. Left atrial and appendage mechanical function after pharmacological or electrical cardioversion in patients with

63. 64.

65. 66. 67.

left atrial appendage occlusion

chronic atrial fibrillation: a multicenter, randomized study. Ital Heart J 2000;1(2):128–​36. Nietlispach F, et al. Amplatzer left atrial appendage occlusion: single center 10-​year experience. Catheter Cardiovasc Interv 2013;82(2):283–​9. Li X, et al. Over one year efficacy and safety of left atrial appendage occlusion versus novel oral anticoagulants for stroke prevention in atrial fibrillation: a systematic review and meta-​analysis of randomized controlled trials and observational studies. Heart Rhythm 2016;13:1203–​14. Masson JB, et al. Transcatheter left atrial appendage closure using intracardiac echocardiographic guidance from the left atrium. Can J Cardiol 2015;31(12):1497, e7–​1497, e14. Perk G, et al. Catheter-​based left atrial appendage occlusion procedure: role of echocardiography. Eur Heart J Cardiovasc Imaging 2012;13(2):132–​8. Pison L, et al. Left atrial appendage closure—​indications, techniques, and outcomes: results of the European Heart Rhythm Association Survey. Europace 2015;17(4):642–​6.

713

714

715

SECTION 9

The Future

48 Novel device therapies for resistant hypertension   717 Kenneth Chan, Manish Saxena, and Melvin D. Lobo 49 Robotic percutaneous coronary intervention   731 Giora Weisz

50 Stem cell delivery and therapy   736 Fizzah Choudry and Anthony Mathur

716

71

CHAPTER 48

Novel device therapies for resistant hypertension Kenneth Chan, Manish Saxena, and Melvin D. Lobo

Introduction—​prevalence, definition, and problem of resistant hypertension Resistant hypertension (RHTN) is defined as uncontrolled office blood pressure (BP) (>140/​90 mmHg) despite treatment with maximum tolerated doses of three or more antihypertensive agents from at least three different classes, including a diuretic (1). The prevalence of RHTN is about 8–​18% in hypertensive patients (2, 3) and confers greatly increased risk of cardiovascular morbidity and mortality (1, 4–​6). An important cause of RHTN is non-​adherence to antihypertensive medications; this can occur for a variety of reasons, including patient-​ related factors and adverse effects of drugs (7). It has been demonstrated that only 50% of newly diagnosed hypertensive patients take their medication 6 months after initial prescription and up to one-​quarter of patients with RHTN do not take any of their antihypertensive drugs (8, 9). As a result, novel treatment strategies are urgently needed to improve BP control in the population with RHTN which could potentially be extended to the wider population of patients with hypertension who may wish to avoid lifelong medications. Incontrovertible evidence indicates that the sympathetic nervous system (SNS) is a key player in all stages of hypertension and that the degree of SNS activation parallels the severity of the hypertensive state (10, 11). Here we discuss four novel device therapies (renal denervation, baroreflex activation therapy, carotid body ablation, and central iliac ateriovenous anastomosis) that have recently been developed for the treatment of hypertension; each therapy addresses separate aspects of the SNS regulation of BP control. In addition, we consider an alternative new BP-​lowering technology, which creates a calibrated central iliac arteriovenous anastomosis and predominantly targets mechanical aspects of the circulation.

Renal sympathetic denervation Rationale and mechanisms of action The renal SNS is comprised of a dense network of afferent and efferent fibres that lie primarily in the adventitia and travel together along the course of the renal artery. Afferent renal sympathetic nerves are activated in response to reflexive stretching of mechanoreceptors, as well as stimulation of chemoreceptors by renal ischaemia and biochemical changes (12, 13). Sympathetic efferent

nerve signalling leads to increased renin secretion by the juxtaglomerular apparatus, renal vasoconstriction, and enhanced sodium reabsorption by epithelial tubular cells (10, 13). A substantial body of data underpins a critical role for renal sympathetic efferent and afferent nerves in the initiation and maintenance of hypertension in animal models and in humans (14). More recently, it has been demonstrated (in several diverse animal models of experimental hypertension) that surgical renal denervation effectively reduced BP (or prevented the development of hypertension) without compromising other kidney functions such as renal bloodflow and glomerular filtration rate (12, 15). Even before this knowledge accrued, in the 1930s and 1940s radical thoracolumbar sympathectomy in patients with severe uncontrolled hypertension (resulting in renal denervation) led to sustained BP lowering and reduced mortality (16–​18). However, this approach was severely limited by disabling complications and was eventually superseded with the introduction of the first antihypertensive drugs in the mid 1950s. Subsequently, observations from patients undergoing bilateral nephrectomy in end-​stage renal disease have demonstrated normalization of central sympathetic outflow, with consistent reduction in BP and systemic vascular resistance suggesting therapeutic utility for renal sympathectomy in humans (19). Techniques such as surgical ligation and reanastomosis of the renal artery used in animal models are not feasible in humans. However, selective renal sympathetic denervation in humans can now be achieved by targeting sympathetic nerves from within the renal artery wall via a percutaneous approach.

Procedure—​device types Current endovascular catheter systems utilize 5–​ 7  F catheters to access renal arteries via the femoral artery. The idea is that radiofrequency (RF) or ultrasound energy (US) delivered by the catheter results in focal frictional heating of the arterial wall, causing destruction of adventitial renal nerves (Fig 48.1). Most studies were done with the unipolar Symplicity/​Flex catheter (Medtronic, Minneapolis, MN, USA), which delivers from four to seven discrete 8-​watt radiofrequency ablations of 120 seconds’ duration to each renal artery. During ablation, the catheter system monitors tip temperature and impedance, delivering energy to a predetermined proprietary algorithm. Patients are usually treated under conscious sedation with titration of opiate analgesia.

718

718

Section 9  the future

Figure 48.1  Schematic cross-​sectional microanatomy of renal artery. Ablation catheter designs: A) first-​generation Symplicity™; B) second-​generation Symplicity Spyral™; C) St. Jude Medical multielectrode EnligHTN™; D) Covidien One-​Shot™ spiral radiofrequency catheter; E) Biosense irrigated multielectrode ThermoCOOL™; F) Verve Medical™ retroureteric multielectrode NephroBlate™; G) Boston Scientific Vessix™ multielectrode, balloon-​mounted bipolar; H) ReCor Medical Paradise™ circumferential irrigated balloon; I) CardioSonic TIVUS™ balloon US catheter; J) Mercator Bullfrog™ microneedle catheter for perivascular guanethidine injection.

Variations upon this theme aim to improve the dependability of focal ablation delivery and make use of multielectrode RF catheters and irrigated RF/​US balloon catheters to minimize endothelial injury (Table 48.1) (20). In addition, catheters are now available for microinjection of neurotoxin (e.g. alcohol) into the adventitial space, and these have the potential to avoid endothelial damage, although inadvertent dispersal of neurotoxin beyond the adventitia into the retroperitoneal space remains a possible concern (21). A  novel RF catheter system (NephroBlate™, Verve Medical, CA, USA) can be introduced transurethrally to ablate within the renal pelvis, which is believed to be densely innervated with afferent fibres (22). This non-​vascular platform has the benefit of overcoming limitations of the endoluminal approach in renal arteries that are anatomically unsuitable for renal denervation (RDN) catheters or patients with bleeding risks.

Clinical trial evidence—​efficacy and safety Symplicity HTN-​1 and HTN-​2 and early studies The proof of concept open-​label Symplicity HTN-​1 study (23) and the randomized controlled open-​label Symplicity HTN-​2 study (24) showed that RF RDN led to striking and highly significant office BP (OBP) reduction of 22/​10  mmHg and 32/​12  mmHg, respectively, at 6  months’ follow-​ up in patients with RHTN (Fig 48.2). Renal noradrenaline spillover (renal efferent activity) and total body noradrenaline spillover (central sympathetic drive via the renal afferent) and muscle sympathetic nerve activity (MSNA) were all reduced after renal denervation (25). A  subsequent meta-​analysis of randomized controlled trials (RCTs) and smaller observational studies showed substantial OBP reduction of

25/​10 mmHg at 6 months, and that there were no significant differences among various catheters (26). However, RDN has still not been proven to be effective in the setting of a sham-​controlled RCT with ambulatory BP endpoints.

Symplicity HTN-​3 In early 2014 the publication of the single blind randomized sham-​ controlled Symplicity HTN-​3 study report came as something of a shock to the device industry and overenthusiastic proponents of RDN because of neutral findings (27). In this study, designed to address criticisms of prior studies, 535 patients were randomized in a 2:1 ratio to receive RDN or renal angiography (sham control), respectively, with patients demonstrably blinded to their treatment allocation. At 6  months, although the RDN group showed significant OBP reduction from baseline (14/​7  mmHg), similar reductions were also observed in the control group (12/​5 mmHg), and thus the study failed to meet the primary efficacy endpoint. Similarly, there were no significant differences in 24-​h ambulatory blood pressure (ABP) lowering, a secondary efficacy endpoint, between the groups (27). There has been much speculation over the failure of HTN-​3 to achieve comparable BP reductions to that demonstrated in previous RDN trials, although attenuated OBP reduction in the setting of an unbiased study was predicted by some (28). However, it remains an inescapable fact that Symplicity HTN-​3 was plagued by critical design and execution flaws that included 40% of patients in both arms undergoing medication changes throughout the study, confounding interpretation of the final results (29, 30). Furthermore, subsequent analysis of the RDN procedures in the study revealed

719

Chapter 48 

novel device therapies for resistant hypertension

Table 48.1  Current renal denervation catheter technologies/​platforms Device (company, CE Mark)

Catheter design

Technical details

Patient criteria

Clinical evidence

Symplicity™ Medtronic Inc. Minneapolis, USA CE: Feb 2008

6 Fr, no guide wire Unipolar, monoelectrode

8 watts, 2 min/​ablation >4 ablation/​artery Cooling: bloodflow

Renal artery >4 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), BP stable on maximum dose of ≥3 meds for 6 m

Feasibility HTN-​1 study (NCT00664638): sustained OBP reduction at 36 m, response rate 87% Randomized controlled HTN-​2 study (NCT00888433): significant OBP reduction at 6 m (32/​12 mmHg) as well as ABP, response rate 84% Double-​blind, sham-​controlled HTN-​3 study (NCT01418261): failed primary efficacy 6 m OBP endpoint, response rate 58.3% Adverse events: renal artery stenosis reported

Symplicity Spyral™ Medtronic Inc. Minneapolis, USA CE: n/​a

6 Fr, over-​the-​wire Unipolar, 4-​electrode Catheter helically conforms to artery

8 watts 1 min/​ablation 1 ablation/​artery Cooling: bloodflow

Renal artery 3–​8 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM)

Observational SPYRAL HTN study (NCT02439775): interim results: significant OBP reduction at 1 m (16/​7 mmHg)a Adverse events: none reported

EnligHTN™ St. Jude Medical St. Paul, USA CE: Dec 2011

8 Fr, no guide wire Unipolar, 4-​electrode Basket mounted (6 or 8 mm)

6 watts 90 s/​sequential ablations per electrode 2 ablation/​artery Cooling: bloodflow

Renal artery 4–​8 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), BP stable on maximum dose of ≥3 meds for 2 w

Observational EnligHTN-​1 study (NCT01438229): significant OBP reduction at 6 m (26/​10 mmHg), response rate 80%b Adverse events: none reported. No renal artery stenosis at 6 m CTA checkb

ThermoCOOL™ Biosense Webster Diamond Bar, USA CE: May 2012

7 Fr, over-​the-​wire Unipolar, 4-​electrode

10–​20 watts 30 s/​ablation 4–​6 ablations/​artery Cooling: irrigated (6-​hole) catheter tip

Renal artery >4 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), BP stable on maximum dose of ≥3 meds for 6 m

RENABLATE study (NCT01756300) and RENABLATE II study (NCT02095691): results pending Adverse events: none reported

Vessix V2™ Boston Scientific Natick, USA CE: May 2012

8 Fr, over-​the-​wire Bipolar, balloon-​mounted, 8-​electrode pairs Non-​compliant balloon (3 atm, 4–​7 mm)

1 watt 30 s/​ablation 1–​2 ablations/​artery Cooling: bloodflow

Renal artery 3–​7 mm Baseline systolic OBP >150 mmHg ABP >135 mmHg eGFR ≥45 ml/​min/​1.73m2

Reduce HTN:REINFORCE study (NCT01541865) and Reduce HTN study (NCT02392351): systolic SBP reduced by 27 mmHg at 6 m, response rate 85% Adverse events: increased small nerve growth at 6 m in preclinical porcine model, unknown significancec

Iberis™ Terumo Corp. Tokyo, Japan CE: Apr2013

4 Fr, no guide wire Unipolar, monoelectrode Radial access

8 watts 2 min/​ablation >4 ablations/​artery Cooling: bloodflow

Renal artery >4 mm Iberis-​HTN Registry (NCT02295683) Baseline systolic BP Case report: OBP reduced by 15/​10 mmHg >160 mmHg (>150 mmHg for (n = 1) 2 w post RSDd T2DM), BP stable on maximum dose of ≥3 meds for 8 w

Nephroblate™ Verve Medical Santa Barbara, USA CE: n/​a

9 Fr, over-​the-​wire Unipolar, multielectrode Retroureteric No systemic contrast

Low power 1 ablation/​renal pelvis Cooling: urinary flow

Baseline systolic BP Early clinical data on patients (n = 3) >160 mmHg (>150 mmHg for undergoing nephrectomye T2DM), BP stable on maximum Adverse effect: none reported dose of ≥3 meds

6 Fr, over-​the-​wire Cylindrical transducer Transducer in low-​ pressure, 5–​8 mm cylindrical, centring balloon

25–​30 watts, non-​focused US Circumferential 40 s/​ablation; 4 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), BP stable on maximum dose of ≥3 meds

Radiofrequency

Ultrasound Paradise™ ReCor Medical Menlo Park, USA CE: Dec 2011

Observational study (n = 11): OBP reduced by 36/​17 mmHg at 3mf REALISE study (NCT01529372): results pending ACHIEVE study (NCT01789918): results pending

(Continued)

719

720

720

Section 9  the future

Table 48.1  (Continued) Device (company, CE Mark)

Catheter design

Technical details

Patient criteria

Clinical evidence

Sound 360™ Sound Interventions Stony Brook, USA CE: n/​a

8 Fr, no guide wire Cylindrical transducer Transducer in low-​ pressure, triangular, centring balloon

Low power, high-​intensity, non-​focused US Circumferential 2 min/​ablation; 2 ablations/​artery Cooling: bloodflow

Renal artery: >5 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), BP stable on maximum dose of ≥3 meds

Observational Sound-​ITV study (n = 10): OBP reduced by 31/​10 mmHg at 1 mg Adverse events: none reported. Post-​RD angiography and IVUS showed no change arterial sizeg

Surround Sound™ Kona Medical Campbell, USA CE: n/​a

External applied US energy

Low-​intensity, focused US 3 min/​ablation; 1 ablation/​artery Cooling: n/​a

Renal artery: >4 mm Baseline systolic BP >160 mmHg, BP stable on maximum dose of ≥3 meds eGFR >45 ml/​min

Sham controlled WAVE IV (NCT02029885): results pending Observational WAVE II study (n = 13): OBP reduced by 18/​0 mmHg at 6 wh Adverse events: none reported

Bullfrog™ Mercator MedSystems Inc. San Leandro, USA CE: Feb 2013

6 Fr, over-​the-​wire

Balloon-​sheathed microneedle (30 G) Balloon inflation (2 atm) exposes microneedle Periadventitial delivery of guanethidine 50 mg in 6 ml

Renal artery: 2–​6 mm

Preclinical: porcine (n = 15) to 28 d post RSD Renal nerve rarefaction, reduced renal NE level. No renal artery morphological changes. Undetectable plasma (guanethidine) 1 d post RSD, 54 ng guanethidine/​g renal tissue 28 d post RSDi

Peregrine™ Ablative Solutions Kalamazoo, USA CE: n/​a

7 Fr, over-​the-​wire 3 guide-​tubes to centre catheter

Tube-​sheathed microneedles (3× 32 G) Periadventitial delivery of 0.15–​0.6 ml dehydrated ethanol

Renal artery: 5–​7 mm Baseline systolic BP >160 mmHg (>150 mmHg for T2DM), stable BP with >2 meds for 4 w

Safety/​efficacy Peregrine study (NCT02155790): results pending Preclinical: porcine (n = 12) to 14–​45 d post RSD. Dose-​dependent reduction in renal NE levels (14 d). Histological confirmation of neural injury. Angiography normal (45 d) post RSDj

40–​80 nm magnetic nanoparticles Internal/​external magnetic field steer particles to renal artery wall Modulation of magnetic field releases Botox®

Renal artery: not known

No published data

Chemical

ApexNano™ Standard catheter ApexNano Therapeutics Herzliya, Israel CE: n/​a

ABP, Ambulatory blood pressure; atm, atmosphere; CTA, computed tomography angiogram; d, day; Egfr, estimated glomerular filtration rate; IVUS, intravascular ultrasound; m, month; meds, antihypertensive medications; NCT, ClinicalTrial.gov identifier number; NE, norepinephrine; OBP, office blood pressure; RSD, renal sympathetic denervation; RAS, renal artery stenosis; RHTN, resistant hypertension; SNS, sympathetic nervous system; T2DM, type 2 diabetes mellitus; US, ultrasound; w, week. Additional references: a Whitbourn R, Harding S, Rothman M, et al. Renal artery denervation with a new simultaneous multielectrode catheter for treatment of resistant hypertension: results from the Symplicity

Spyral first-​in-​man study (abstract). J Am Coll Cardiol 2013;62:B150. b Worthley SG, Tsioufis CP, Worthley MI, et al. Safety and efficacy of a multi-​electrode renal sympathetic denervation system in resistant hypertension: the EnligHTN I trial. Eur Heart J

2013;34:2132–​40. c Mazor M, Baird R, Stanley J. Evaluation of acute, sub-​acute, and chronic renal artery nerve morphological changes following bipolar radiofrequency renal denervation treatment in the

porcine model (abstract). J Am Coll Cardiol 2013;62:B150. d Honton B, Pathak A, Sauguet A, et al. First report of transradial renal denervation with the dedicated radiofrequency Iberis™ catheter. EuroIntervention 2014;9(12):1385–​8. e Heuser R, Buelna T, Gold A, et al. Preclinical and early clinical experience of a non-​vascular treatment for resistant hypertension. J Am Coll Cardiol 2014;64(11_​S): doi:10.1016/​

j.jacc.2014.07.459 f Mabin T, Sapoval M, Cabane V, et al. First experience with endovascular ultrasound renal denervation for the treatment of resistant hypertension. EuroIntervention 2012;8:57–​61. g Neuzil P, Petru J, Vondrakova D, et al. Circumferential therapeutic ultrasound for the treatment of resistant hypertension: preliminary results of human feasibility study (SOUND-​ITV)

(abstract). J Am Coll Cardiol 2012;60:B101–​2. h Neuzil P, Whitbourn RJ, Starek Z, et al. Optimized external focused ultrasound for renal sympathetic denervation—​Wave II trial (abstract). J Am Coll Cardiol 2013;62:B20. i Owens CD, Gasper WJ, Rousselle S, et al. Peri-​adventitial renal artery delivery of guanethidine monosulfate attenuates renal nerve function: preclinical experience and implication for

resistant hypertension (abstract). J Am Coll Cardiol 2011;58:B120. j Fischell TA, Vega F, Raju N, et al. Ethanol-​mediated perivascular renal sympathetic denervation: preclinical validation of safety and efficacy in a porcine model. EuroIntervention

2013;9:140–​7.

721

Mean systolic blood pressure change at 6-month (mmHg)

Chapter 48 

using spironolactone (32). This study was strengthened by optimized RHTN work-​up by accredited specialists and confirmed medication adherence using plasma drug assays. Notably, 39% of patients in the control group discontinued therapy with spironolactone during the study due to intolerance. A recent meta-​analysis of RCTs of RDN involving 985 patients, of whom 588 were treated with the Symplicity™ system, showed heterogeneous effects for OBP and ABP outcomes, with no overall significant decrease in BP; however, it confirmed the safety of the procedure (33). Fig 48.3 shows a summary of BP changes in clinical trials to date.

10 0 –10 –20 –30

n = 353/171 n = 45

–40 Symplicity HTN-1

n = 52/54 Symplicity HTN-2

Renal denervation

novel device therapies for resistant hypertension

Symplicity HTN-3

Registry data

Control

Figure 48.2  Blood pressure responses in Symplicity HTN trials. Mean changes in office systolic blood pressure at 6-​month follow-​up reported in the Symplicity trials. Error bars represent 95% confidence intervals (24, 25, 28). Data sourced from Kapil et al. Renal sympathetic denervation—​a review of applications in current practice. Interventional Cardiology Review, 2014;9(1):54–​61, and Heuser RR, et al. A novel non-​vascular system to treat resistant hypertension. EuroIntervention, 2013;9(1):135–​9.

that only 5% of patients received per protocol ablation therapy and that the 19 patients who received bilateral RDN in all four quadrants exhibited the greatest reductions in office, home and ambulatory systolic blood pressure (SBP) (–​24.3, –​9.0, and –​10.3 mmHg, respectively) (29).

Other randomized controlled trials of renal denervation Data from other recent RCTs indicate that, when performed properly, RDN can be effective. In the DENER-​HTN study, funded by the French Ministry of Health, investigators compared the efficacy of RF RDN (using the Symplicity catheter) in 48 RHTN patients taking standardized optimized antihypertensive therapy (SOAT) to 53 patients taking SOAT alone. The findings were that RDN plus SOAT lowered daytime ambulatory SBP by 15.8  mmHg versus 9.9 mmHg in the control group (P = 0.0329) with minimal adverse events. The merits of this study included being carried out in tertiary referral centres with interventionalists who had substantial prior experience of RDN and incorporating a meticulously executed medication regimen that ensured similar antihypertensive therapy and adherence between the groups at 6 months (31). In the Prague-​15 study, 52 RHTN patients who had RDN experienced comparable 24-​h average SBP reduction were compared to 54 control group patients managed with intensified pharmacotherapy

The prospective Global SYMPLICITY Registry showed that, in 998 patients with baseline office SBP of 163.5 mmHg and baseline ambulatory SBP of 151.5 mmHg, office SBP and ambulatory SBP were reduced by 11.6 mmHg and 6.6 mmHg, respectively, at 6 months (P < 0.001 for both) (34). In a cohort with more severe hypertension (baseline office SBP/​ambulatory SBP 179.3/​159.0 mmHg), office SBP and ambulatory SBP were reduced by 20.3 and 8.9 mmHg, respectively (P < 0.001 for both). In this multicentre study RDN was associated with low rates of adverse events. The single-​centre ALSTER and Heidelberg registries also showed response rates of 76% (n = 93) and 73% (n = 63), respectively (35, 36). Taken together, these data support the effectiveness of RDN in the real world and stand in contrast to the results from RCTs.

How safe is renal denervation? The clinical trials to date demonstrate that RDN is a relatively safe procedure with a major adverse event rate of about 1.4% against an objective performance criterion of 50%

Improved echocardiographic parameters of LV hypertrophy and stiffness, 17% reduction in LV mass at 6 months, and improved diastolic relaxation and systolic function

Mahfoud et al. (94)

RDN versus OMT (n = 72)

Baseline mean BP 170/​ 90 mmHg, EF >50%,

LV mass index reduced by 7.1% at 6 months assessed by magnetic resonance imaging. Improved systolic function

Pokushalov et al. (95)

Pulmonary vein isolation (PVI) +/​-​RDN (n = 14/​13)

Baseline mean BP 181/​ 97 mmHg

PVI+RDN reduce AF relapse for 12 months (69% versus 29%)

McLellan et al. (96)

Single arm n = 14

Baseline mean ambulatory BP 152/​84 mmHg

BP reduction after RDN improves regional and global atrial conduction and reduces ventricular mass/​fibrosis

Pokushalov et al. (97)

2-​cohort meta-​analysis n = 86

Moderate to severe hypertension

Hazard ratio for PVI+RDN versus PVI alone = 0.25

Mahfoud et al. (98)

Single RDN arm n = 50

Resistant hypertension cohort from HTN-​1 trial

Improved fasting glucose, serum insulin, and C-​peptide levels at 3 months

DREAMs study (99)

Single RDN arm n = 29

Metabolic syndrome on ≤1 antihypertensive or antidiabetic

No significant improvement of insulin sensitivity at 12 months after RDN

Witkowski et al. (100)

Single RDN arm n = 10

Moderate–​severe OSA

Reduction in post-​oral glucose tolerance hyperglycaemia and HbA1c at 6 months

Schlaich et al. (101)

Single RDN arm n = 2

Polycystic ovary syndrome

Improvement in insulin sensitivity by 17.5% in absence of weight changes at 3 months

Hering et al. (102)

RDN single arm n = 15

eGFR 45 ml/​min/​1.73 m2 macroalbuminuria at 6 months

Witkowski et al. (100)

Single RDN arm n = 10

Moderate–​severe OSA

Reduce apnoea–​hypopnoea index and subjective sleepiness

Zhao et al. (106)

RDN versus CPAP n = 31

Moderate–​severe OSA

Improved nocturnal apnoea–​hypopnoea index but more significant improvement in CPAP group

Systolic heart failure

Heart failure with preserved ejection fraction

Atrial fibrillation

Glycaemic control

Chronic kidney disease

Obstructive sleep apnoea

AF, Atrial fibrillation; BP, blood pressure; CKD, chronic kidney disease; CPAP, continuous positive airway pressure; EF, ejection fraction; GFR, glomerular filtration rate; LV, left ventricular; NYHA, New York Heart Association classification; NT-​pro BNP, NT-​pro brain natriuretic peptide; OMT, optimized medical therapy; OSA, obstructive sleep apnoea; RDN, renal denervation.

723

724

724

Section 9  the future

Carotid Chemoreceptors

Sympathetic tone

Carotid Baroreceptors

Sympathetic tone

bipolar electrodes, which were attached to the carotid sinus under general anaesthesia (Fig 48.5). The use of electrical field stimulation instead of direct activation was intended to reduce stimulation of surrounding muscles and mitigate the risk of poor electrode contract over time (64). Subsequently, a second-​generation device (Barostim Neo™) encompassing a single unipolar electrode has become available, with improved battery life from a miniaturized generator (70) (Fig 48.5). The electrode is usually placed on the right carotid sinus and the procedure can now be done under conscious sedation.

Clinical trial data Vasoconstriction

HR Contractility

RBF/GFR Renin Na+/Volume

Figure 48.4  Schematic of carotid chemoreceptor and baroreceptor reflexes. CC, Common carotid; EC, external carotid; IC, internal carotid; GFR, glomerular filtration rate; HR, heart rate; Na+, sodium; RBF, renal bloodflow. Carotid body image adapted from J Paton et al, Hypertension, 2013;61:5–​13.

Procedure Baroreflex activation therapy (BAT) now uses novel technology to deliver electrical field stimulation at the carotid sinus, thus reducing elevated BP. The first-​generation Rheos® system (no longer available) consisted of an implantable pulse generator (IPG) and utilized

The Rheos device was initially evaluated in a feasibility study:  the non-​randomized, open-​label Device-​Based Therapy of Hypertension Trial (DEBuT-​HT) in 45 patients with RHTN, and showed an average BP reduction of 21/​12 mmHg at 3 months and 33/​22 mmHg reduction at 2 years (71). Subsequently, in the Rheos Pivotal Trial (72), 265 patients were randomized in a 2:1 fashion to early (1-​month post-​implantation) or delayed device activation (6 months post-​implantation). Although 42% of participants in the early group versus 24% of the delayed group achieved SBP 15,000 published cases of CB resection (76).

Clinical trial data and future perspectives A proof-​of-​concept study of unilateral CB ablation in patients with RHTN has demonstrated significant and sustained office BP reduction of 23/​12  mmHg at 6  months postoperatively in 8 out of 15 patients who had evidence of increased baseline CB activity. No serious adverse events were observed at up to 12  months of follow-​up, and hypoxic ventilatory drive was maintained (80). The feasibility of unilateral endovascular CB ablation using the Cibiem Carotid Body Modulation System  is currently being evaluated (ClinicalTrials.gov: NCT02099851).

novel device therapies for resistant hypertension

Central iliac arteriovenous anastomosis Rationale and mechanism of action The devices discussed earlier all target different aspects of the SNS regulation of BP. However, a novel approach that has recently come to light focuses predominantly upon addressing mechanical aspects of the circulation. By creating a conduit between the proximal arterial and low resistance venous circulation, the central iliac arteriovenous (AV) anastomosis provides a unique opportunity for improving proximal vascular compliance, which helps to restore at least part of the Windkessel effect (81–​83). Furthermore, reducing effective arterial volume can lower systemic vascular resistance, thereby reducing cardiac afterload and BP. In addition, it is likely that, by increasing venous oxygenation and increasing right heart stretch through increased preload, the device has some sympathomodulatory effects that also contribute to BP reduction (81).

Procedure The ROX AV coupler is a nitinol stent-​like device that creates a 4-​mm anastomosis between the external iliac artery and vein in a procedure undertaken in a standard cardiac catheterization laboratory under fluoroscopic guidance (84). An advantage of the ROX coupler system is that, unlike RDN procedures, deployment is verifiable and reversible if required. The resultant diversion of a calibrated amount of arterial blood (0.8–​1 l/​min) into the proximal large capacitance venous circuit produces an immediate reduction in both SBP and diastolic BP and obviates any contribution from placebo/​Hawthorne effects (81, 85) (Fig 48.6).

Clinical data and safety considerations Therapeutic use of a central iliac AV anastomosis was initially studied in patients with severe chronic obstructive pulmonary disease (COPD), and led to improved 6-​min walking distance at 6–​12 weeks following the procedure (86). Importantly, no adverse effects were reported. An open-​label study of 24 patients with COPD and mild hypertension subsequently demonstrated a reduction in OBP from 145/​86 to 132/​67  mmHg (P < 0.001) at 12  months with no change in medications compared to baseline (87). In the randomized controlled open-​label ROX CONTROL HTN trial (88), 83 patients in total received standard care or insertion of AV coupler with standard care. At 6 months, OBP and ABP were reduced by 27/​20 and 14/​14  mmHg, respectively, in the coupler group (P < 0.0001 for all changes), whilst in the control group there was no significant change in either (Fig 48.7). In a small subset of patients who had prior RDN, there was highly significant OBP reduction (34/​22 mmHg) and ABP reduction (12/​15) in the coupler group, with no significant change in the control group for either, suggesting that AV coupler therapy may be beneficial in cases where sympathomodulation has failed (89). The main complication reported was ipsilateral venous stenosis in 29% of the patients in the coupler group, which was managed by venoplasty and/​or stenting in all patients to good effect. BP differences between the coupler and control group did not reflect differences in use of medication and there was a reduction in hospitalizations for hypertensive urgencies in the coupler group (88). Thus far there have been no reports of high output cardiac failure with this device.

725

726

Section 9  the future

A

B

Figure 48.6  A) ROX device and placement under fluoroscopic guidance. B) Immediate, variable blood pressure reduction response. Images reproduced with permission from ROX Medical.

Future directions At present very few patients have experienced AV coupler therapy for hypertension worldwide and there is much to be learnt about the device and its long-​term effects. It remains to be seen whether the device/​procedure can be optimized to reduce the incidence of venous stenosis by minimizing turbulence from the arterial inflow into the venous segment. Furthermore, it may in future be possible to vary coupler size to ‘start small’ for those patients in whom it may be more risky to increase cardiac output.

For the time being there are still concerns regarding long-​term safety, particularly in respect of the potential for development of high output cardiac states, although with a fixed calibre anastomosis the risks of this should be minimal in the absence of comorbidities such as anaemia/​thyrotoxicosis/​septicaemia. It is therefore reassuring that the coupler device is fully reversible with a covered stent and that, to date, there are no reports of high output cardiac failure in treated patients. Ongoing evaluation of the therapy is taking place within a global registry study (ClinicalTrials.gov: NCT1885390) (85). A US-​based

AV Coupler Group

30

OBP n = 42

Control Group

24-h ABP n = 42

OBP n = 34

24-h ABP n = 35

20 Change from baseline to 6 mo (mmHg)

726

10 0 –10 –20 –30 –20.1 p