hemodynamic measurements in coronary, valvular, and ... · hemodynamic measurements in coronary,...

Hemodynamic measurements in coronary, valvular, andperipheral vascular disease : the role of the medical engineerin a cardiovascular department of a non-academic heartcenterCitation for published version (APA):Veer, van 't, M. (2008). Hemodynamic measurements in coronary, valvular, and peripheral vascular disease :the role of the medical engineer in a cardiovascular department of a non-academic heart center. Eindhoven:Technische Universiteit Eindhoven. https://doi.org/10.6100/IR638545

DOI:10.6100/IR638545

Document status and date:Published: 01/01/2008

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 03. Apr. 2020

https://doi.org/10.6100/IR638545

https://doi.org/10.6100/IR638545

https://research.tue.nl/en/publications/hemodynamic-measurements-in-coronary-valvular-and-peripheral-vascular-disease--the-role-of-the-medical-engineer-in-a-cardiovascular-department-of-a-nonacademic-heart-center(28465dac-6d79-43e5-8b4f-2db1772c9466).html

Hemodynamic measurementsin coronary, valvular, and

peripheral vascular disease

The role of the Medical Engineer in a cardiovasculardepartment of a non-academic heart center

A catalogue record is available from the Eindhoven University of Technology Library

ISBN: 978-90-386-1439-7

Copyright c©2008 by M. van ’t Veer

All rights reserved. No part of this book may be reproduced, stored in a database orretrieval system, or published, in any form or in any way, electronically, mechanically,by print, photoprint, microfilm, or any other means without prior written permissionof the author.

Cover design by Oranje Vormgevers, Eindhoven, The Netherlands.After the idea of Mariska van ’t Veer, Wageningen, The Netherlands

Printed by Drukkerij De Budelse, Budel, The Netherlands.

The research described in this thesis was financially supported by educationalgrants of Stichting Vrienden van het Hart, Eindhoven, The Netherlands; RADIMedical Systems, Uppsala, Sweden; Medtronic, Heerlen, The Netherlands ; Cordis,Roden, The Netherlands; and of the Wetenschappelijk Fonds Catharina Ziekenhuis,Eindhoven, The Netherlands.



The role of the Medical Engineer in a cardiovasculardepartment of a non-academic heart center

PROEFSCHRIFT

ter verkrijging van de graad van doctoraan de Technische Universiteit Eindhoven,

op gezag van de Rector Magnificus, prof.dr.ir. C.J. van Duijn,voor een commissie aangewezen door het College voor Promoties

in het openbaar te verdedigen opdonderdag 18 december 2008 om 16.00 uur

door

Marcel van ’t Veer

geboren te Soest

Dit proefschrift is goedgekeurd door de promotoren:

prof.dr. N.H.J. Pijlsenprof.dr.ir. F.N. van de Vosse

Copromotor:dr.ir. M.C.M. Rutten

v

Men zegt dat geluk geen meervoud heeft.Ik vind van wel.

voor pa en ma

Contents

1 Introduction 11.1 Hemodynamic measurements in the

cardiovascular system . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 The Medical Engineer . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Outline of this thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Degenerative diseases of the cardiovascular system 52.1 Coronary artery disease . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.1.1 Pathophysiology . . . . . . . . . . . . . . . . . . . . . . . . . . 52.1.2 Anatomic assessment of the coronary circulation . . . . . . . . 62.1.3 Physiologic assessment of the coronary circulation . . . . . . . 62.1.4 Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.2 Valvular disease. The aortic valve . . . . . . . . . . . . . . . . . . . . . 102.2.1 Clinical assessment . . . . . . . . . . . . . . . . . . . . . . . . . 112.2.2 Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.3 Peripheral vascular disease. The Abdominal Aortic Aneurysm . . . . . 132.3.1 Anatomic assessment. The diameter criterium . . . . . . . . . . 132.3.2 Physiologic assessment. Hemodynamic measurements . . . . . 14

3 Hemodynamic evaluation of coronary stents 153.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.2.1 Study population . . . . . . . . . . . . . . . . . . . . . . . . . . 163.2.2 Interventional protocol . . . . . . . . . . . . . . . . . . . . . . . 183.2.3 Hemodynamic analysis . . . . . . . . . . . . . . . . . . . . . . . 183.2.4 Angiographic analysis . . . . . . . . . . . . . . . . . . . . . . . 193.2.5 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.3.1 Baseline characteristics and procedural results . . . . . . . . . . 203.3.2 Clinical follow-up at 6 months . . . . . . . . . . . . . . . . . . . 203.3.3 Angiographic follow-up at 6 months . . . . . . . . . . . . . . . 223.3.4 Hemodynamic follow-up and WSS at 6 months . . . . . . . . . 22

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

vii

viii Contents

3.4.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

4 Influence of orientation of mechanical bi-leaflet valve prosthesis oncoronary perfusion pressure 314.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.1 Study population . . . . . . . . . . . . . . . . . . . . . . . . . . 334.2.2 Echocardiography . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.3 Surgical procedure . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.4 Catheterisation at follow-up . . . . . . . . . . . . . . . . . . . . 344.2.5 Data analysis and statistical analysis . . . . . . . . . . . . . . . 36

4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364.3.1 Procedural results . . . . . . . . . . . . . . . . . . . . . . . . . 364.3.2 Echocardiographic results . . . . . . . . . . . . . . . . . . . . . 384.3.3 Hemodynamic measurements . . . . . . . . . . . . . . . . . . . 38

4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404.4.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Continuous infusion thermodilution for assessment of coronary flow:Theoretical background and in-vitro validation 435.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5.2.1 Theoretical background and measurement principle . . . . . . 455.2.2 In-vitro model and instrumental set-up . . . . . . . . . . . . . . 465.2.3 Measurement protocol . . . . . . . . . . . . . . . . . . . . . . . 47

5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495.3.1 Temperature course of guide wire pullback . . . . . . . . . . . . 495.3.2 Flow measurement at fixed position . . . . . . . . . . . . . . . 49

5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

5.5.1 Clinical implications . . . . . . . . . . . . . . . . . . . . . . . . 54

6 Continuous infusion thermodilution for assessment of coronary flow:Animal study 556.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566.2 Theoretical background and aim . . . . . . . . . . . . . . . . . . . . . 566.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

6.3.1 Animal instrumentation . . . . . . . . . . . . . . . . . . . . . . 576.3.2 Cardiac catheterisation . . . . . . . . . . . . . . . . . . . . . . . 576.3.3 Experimental protocol . . . . . . . . . . . . . . . . . . . . . . . 596.3.4 Measurement procedure . . . . . . . . . . . . . . . . . . . . . . 596.3.5 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . 62

6.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Contents ix

6.4.1 Hemodynamic characteristics and procedural results . . . . . . 626.4.2 Flow measurements . . . . . . . . . . . . . . . . . . . . . . . . 62

6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666.5.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676.5.2 Clinical implications . . . . . . . . . . . . . . . . . . . . . . . . 68

6.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

7 Continuous infusion thermodilution for assessment of coronary flow:Human study 717.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727.2 Theoretical background and aim . . . . . . . . . . . . . . . . . . . . . 727.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

7.3.1 Patient selection . . . . . . . . . . . . . . . . . . . . . . . . . . 727.3.2 Cardiac catheterisation and experimental protocol . . . . . . . 737.3.3 Measurement procedure . . . . . . . . . . . . . . . . . . . . . . 757.3.4 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . 77

7.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 777.4.1 Baseline characteristics and procedural results . . . . . . . . . . 777.4.2 Flow measurements . . . . . . . . . . . . . . . . . . . . . . . . 78

7.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817.5.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827.5.2 Clinical implications . . . . . . . . . . . . . . . . . . . . . . . . 83

7.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

8 Biomechanical properties of abdominal aortic aneurysms assessed bysimultaneously measured pressure and volume changes in humans 858.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 868.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

8.2.1 Study population . . . . . . . . . . . . . . . . . . . . . . . . . . 878.2.2 Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . 878.2.3 Pressure measurement . . . . . . . . . . . . . . . . . . . . . . . 878.2.4 Detection of volume changes . . . . . . . . . . . . . . . . . . . 888.2.5 Biomechanical properties of the AAA . . . . . . . . . . . . . . . 88

8.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 898.3.1 Baseline characteristics and clinical results . . . . . . . . . . . . 898.3.2 Aneurysmal volume change and mechanical properties . . . . . 898.3.3 Invasive and non-invasive blood pressure . . . . . . . . . . . . 90

8.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 938.4.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

9 General discussion and conclusions 979.1 Hemodynamic measurements . . . . . . . . . . . . . . . . . . . . . . . 979.2 The Medical Engineer . . . . . . . . . . . . . . . . . . . . . . . . . . . 1009.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

x Contents

References 102

Summary 113

Samenvatting 117

Dankwoord 121

Curriculum Vitae 123

Publications 125

Chapter 1

Introduction

1.1 Hemodynamic measurements in thecardiovascular system

Development in health care is mainly focussed on improving patient care. Technicalimprovements in this field have led to better methods for diagnosis and treatmentof many diseases. As an example, vast improvements in imaging techniques havecontributed to the ability to analyze anatomic structures in more detail as well as toperform several functional measurements. Furthermore, ongoing developments andminiaturization of sensors enable (invasive) hemodynamic measurements throughoutthe cardiovascular system.

For the cardiovascular system both imaging techniques and hemodynamic mea-surements play a central role in the diagnosis and treatment of a number ofpathologic conditions. One such condition is atherosclerotic coronary artery disease(CAD). Presently, the coronary angiogram is most often used to detect coronaryabnormalities. However, the angiogram only offers an anatomic view of the extent ofthe coronary artery disease. Hence, the functional severity cannot be assessed usingangiography alone. Guide wire mounted sensors have enabled functional assessmentof epicardial coronary artery disease, but challenges remain to assess microvasculardysfunction.

Functional assessment is also embedded in daily practice to assess valvulardisease. Doppler-echocardiography is used to perform physiologic measurementsin addition to anatomic measurements of structures in the heart. This allows thephysician to non-invasively estimate the functional severity of valvular pathologyand to assess the function of a valvular prosthesis at follow-up whenever valvereplacement has taken place. The question whether different methods of aortic valvereplacement have effect on coronary physiology cannot directly be answered usingstandard techniques and requires dedicated hemodynamic measurements.

Comparable to detection of CAD, a purely anatomy-based approach is used in thedetermination of peripheral vascular disease as well. A relatively common vascular

1

2 Chapter 1

disorder is an abdominal aortic aneurysm (AAA). The risk of a AAA is rupture,associated with a high mortality rate. Rupture risk stratification of AAAs is often basedon measurement of maximal diameter alone determined from computed tomography(CT), ultrasound (US), or magnetic resonance imaging (MRI) scans. Such a anatomy-based cut-off value does not take hemodynamic and bio-mechanical parameters intoaccount. To determine these parameters both imaging of the AAA and hemodynamicmeasurements are required.

New methods for diagnosis and treatment improve healthcare but also raisenew research questions. Hemodynamic measurements in combination with standarddiagnostic methods are complementary to answer such questions, specifically in thecardiovascular disease.

1.2 The Medical Engineer

Generally, advanced techniques for acquiring and analyzing data have made healthcare increasingly complex. Medical engineers might have an additive value in theanalyses of the extensive amounts of acquired data using mathematical models to,e.g., understand underlying (pathologic) processes of a disease/phenomenon orpredict the outcome of an intervention. As a consequence, these engineers shouldhave profound knowledge of the biologic and physiologic side of the problem as wellas the capabilities to understand the physical processes and the way these processesare described using mathematical models. From this perspective there is a need formedical engineers that are capable of objectively processing data using mathematicalor computer aided models to support clinical decision making.

The Department of Biomedical Engineering of the Eindhoven University of Tech-nology educates (bio)medical engineers. Besides engineering skills, like planning,performing, and evaluating scientific experiments, a medical engineer learns toquantify and objectify patient data as well. Additionally, model based problem solvingis part of the basic skills.

The training of a medical engineer, however, has special demands for communi-cation, research methods, and interpretation of obtained results. The communicationskills are opportune e.g., towards physicians to translate medical research questionsinto engineering ones (and vice versa). Medical research questions often ensue frompractical issues in daily practice whereas biomedical questions mostly aim to answerresearch questions that deal with the fundamental underlying processes. To be ableto function as a bridge between clinical research and more fundamental science, amedical engineer should have the above mentioned engineering competencies at hisdisposal but also sufficient clinical knowledge.

In performing research at the cutting edge of life sciences and engineering, alsoethical issues play an important role in the choice for research methods. Moreover,especially in a non-academic hospital with high patient load, proper planning ofexperiments is needed to minimize interference with daily practice.

Introduction 3

1.3 Outline of this thesis

This thesis describes a number of research projects that focus on the sometimescomplex background of investigational techniques and accompanying interpretationof hemodynamic measurements. Besides the separate projects that aimed toanswer specific research questions that arose from clinical practice, the overallgoal was to establish the role of the medical engineer in stimulating, initiating,and executing clinical experimental research using hemodynamical measurementsin the cardiovascular system. The projects have been performed at the departmentof Biomedical Engineering of the Eindhoven University of Technology and at thedepartments of cardiology, cardiothoracic surgery, and vascular surgery of theCatharina Hospital in Eindhoven, which is one of the largest cardiovascular centersin Europe. Because hemodynamic measurements in several pathologic conditions arethe main focus in each chapter a general introduction in degenerative diseases of thecardiovascular system is first given in chapter 2.

Chapter 3 describes a study performed at the department of cardiology in whichthe hemodynamic characteristics of two different types of coronary stents werecompared at implantation and at six-month follow-up. So far, only anatomic datahad been available for comparison of these stents.

In cooperation with the department of cardiothoracic surgery, the study describedin chapter 4 was performed to answer the question whether the orientation of amechanical heart valve prosthesis could influence coronary physiology both in restingconditions and during exercise.

Being able to measure absolute coronary blood flow, has been a holy grail incardiology for more than 40 years. Measuring absolute coronary blood flow as well ascoronary pressure simultaneously would mean a great step forward in understandingcoronary physiology. In chapter 5 the validation of a such technique based onthermodilution is described, which was developed at the department of BiomedicalEngineering of the Eindhoven University of Technology. The technique was validatedin-vivo in dogs in the Central Animal Laboratory of the University of Maastricht andsoon thereafter tested and applied in the catheterisation laboratory of the CatharinaHospital Eindhoven. The animal study is described in chapter 6 and the human studyis described in chapter 7.

In chapter 8 a research project is described to estimate global bio-mechanicalproperties of the wall of the aorta in-vivo in patients with an abdominal aorticaneurysm. Simultaneous aneurysmal volume and intra-vascular blood pressuremeasurements were carried out. This study was achieved in close collaboration withthe departments of vascular surgery and radiology.

A general discussion is presented in chapter 9 and the role of the medical engineeris a large heart center is discussed.

4 Chapter 1

Chapter 2

Degenerative diseases of thecardiovascular system

Despite the ability of the human body to adapt to pathologic conditions of thecardiovascular system to a certain extent, catheter based or surgical interventionmight eventually be inevitable. The decision to intervene is often based on anatomicassessment and not on hemodynamic measures to determine the severity of thedisease. A basic overview of different degenerative processes and their hemodynamicconsequences throughout the cardiovascular system is outlined in this chapter.

The most important degenerative process in the human circulation is atherosclero-sis, affecting both large and small arteries and the microvasculature. Atherosclerosishas various manifestations throughout the vascular system. It is generally believedthat atherosclerosis plays an important role in the initiation and the progression ofatherosclerotic plaques. Despite the fact that the first signs of atherosclerosis areobserved in adolescents, the disease is usually slowly progressing and symptomstypically occur a few decades later.

2.1 Coronary artery disease

When atherosclerosis affects the coronary arteries, atherosclerotic plaques causelocal or diffuse narrowings throughout the coronary arteries. As a result of thesenarrowings coronary blood flow can become impaired, which consequently leads tomyocardial ischemia.

2.1.1 Pathophysiology

Myocardial ischemia is the result of an imbalance between the oxygen demand andthe oxygen supply to the heart muscle. The coronary arteries supply the heart withoxygen and nutrients and arise just above the aortic valve (figure 2.1). The epicardialarteries (∅ 2.5-4 mm) branche into smaller arteries and arterioles to eventually form

5

6 Chapter 2

a network of capillaries. Vessels with a diameter >0.4 mm, are defined as conductivevessels or conduit vessels (Pijls and DeBruyne, 2000, chapter 2). Vessel with adiameter between 100 and 400 µm are called pre-arterioles, whereas vessels <100µmare called arterioles. Together these vessels form the resistive vessels (Chilian et al.,1989). The branching continues until thin walled capillaries are formed that perfusethe myocardium(∅ 5 µm). Here the actual exchange of oxygen, nutrients, and wasteproducts takes place.

In a normal coronary system the conductive vessels offer negligible resistanceto blood flow compared to the microcirculation (vessels <200µm). The resistanceof the coronary circulation, here defined as the mean pressure difference over themyocardium divided by the mean myocardial flow, is mainly regulated by the vasculartone of the arterioles. Therefore coronary blood flow is primarily determined byvariations in resistance of these vessels. Depending on the demand of oxygen, valuesof blood flow can increase by a factor 5 during exercise compared to resting (baseline)conditions in healthy humans. This regulatory mechanism can also compensatefor resistances caused by a stenosis in the epicardial vessels. The compensation inarteriolar resistance will, however, be at the cost of maximal achievable flow. Thisexplains why, especially during exercise, patients might experience chest pain as theresult of myocardial ischemia.

2.1.2 Anatomic assessment of the coronary circulation

Usually, patients with typical chest pain eventually undergo a cardiac catheterisation.During cardiac catheterisation a guiding catheter is introduced through whichcontrast dye is injected to visualize the lumen of the coronary artery using X-rayfluoroscopy. These two-dimensional projections of the coronary arteries are assessedvisually or by semi-automatic edge detection algorithms like quantitative coronaryangiography (Reiber et al., 1985). Regardless of the technique used, stenosis severitycan be misjudged for many reasons as is shown in figure 2.2.

Even when a more accurate impression of the anatomic severity is obtainedinvasively, i.e. by intra-vascular ultrasound (IVUS), still no physiologic information isobtained. The functional severity of a stenosis is further influenced by the extent ofthe myocardial perfusion area, the presence of collateral flow, and the resistance ofthe microvascular bed. Non of these factors can be taken into account by anatomicanalysis of the stenosis itself (Gould et al., 1990).

2.1.3 Physiologic assessment of the coronary circulation

Several indices have been proposed to assess coronary artery disease in a physiologicway. An index that describes the vasodilatory reserve of the coronary circulation isthe coronary flow reserve or CFR. CFR is defined as the mean hyperemic coronaryblood flow divided by the mean coronary resting blood flow (Gould et al., 1990).

CFR =Qhyp

Qrest(2.1)

Degenerative diseases of the cardiovascular system 7

Figure 2.1: The right coronary artery (RCA), arising from the right coronary cusp,runs through the atrioventricular groove towards the backside of the heart. TheRCA gives rise to several branches that perfuse the right ventricle, the inferior partof the the interventricular septum, and, in the majority of the cases, part of theposterolateral side of the left ventricular wall as well. The left coronary artery (LCA)arises from the left coronary cusp and usually bifurcates into the circumflex artery(Cx) and the left anterior descending artery (LAD). The LAD gives rise to severalbranches that supply the superior part of the interventricular septum, and the anteriorside of the left ventricular wall. The Cx supplies the lateral side of the left ventricularwall (Zipes et al., 2005, chapter 18).(Medical Illustrations copyright c©1997-2008Nucleus Medical Art,Inc. All Rights Reserved. www.nucleusinc.com)

8 Chapter 2

Figure 2.2: Different projections of schematically drawn coronary stenoses may leadto misjudgement of the true anatomic severity. The black area within the circlerepresents the intact lumen of the artery. On the left the anatomically severelynarrowed coronary artery appears almost normal in the different projections. On theright the coronary artery appears to be severely narrowed in one projection whereasin the other projection the artery look normal. No information of the functionalimportance of an epicardial coronary stenosis is obtained in this way (adapted from(Aarnoudse, 2006))

CFR is influenced by both the epicardial and the microvascular condition of thecoronary system, but does not distinguish between an increased epicardial resistanceor a microvascular flow impairment. Moreover, CFR is dependent on age, heart rate,arterial blood pressure, and has a wide range of normal values.

Fractional flow reserve, or FFR, is an index that describes the hyperemic flow inthe presence of a stenosis (QS

hyp) relatively to the hyperemic flow in the case that thecoronary artery would be completely normal (QN

hyp). By definition the prerequisitefor the determination of FFR is that hyperemia is present. In this case the myocardialresistance (R) is minimal and constant resulting in averaged maximal blood flowduring the heart cycle. Consequently it can be derived that the ratio of QS

hyp andQN

hyp can be described by a ratio of mean pressures. When central venous pressure isassumed to be small (Pv << Pd), FFR can be determined as the mean pressure distalin the coronary artery (Pd) divided by the mean aortic pressure (Pa) both measured atmaximum hyperemia, equation 2.2. Besides the fact that FFR has a uniform normalvalue of 1 for every coronary artery and every person, it is largely independent onarterial pressure, heart rate, or the status of the microcirculation.

FFR =QS

hyp

QNhyp

=(Pd − Pv)/R

(Pa − Pv)/R≈ Pd

Pa(2.2)

Other concepts than CFR and FFR are necessary to describe the specific statusof the microcirculation. For a physiologic description of the microcirculation, meanmicrovascular resistance has been defined as the mean perfusion pressure dividedby the mean myocardial flow. However, it has not been possible to determine


(mean) absolute coronary or myocardial blood flow during catheterisation so far.Therefore an index of microcirculatory resistance, or IMR, has been proposed byFearon et al.(2003) and Aarnoudse et al. (2004b). Instead of absolute values of meanblood flow, a derivative of flow is used: the mean transit time, Tmn, determined usingthermodilution. The mean transit time is inversely proportional to the blood flow andis related to the time a bolus of cold saline needs to travel down the coronary arterywhen injected into the ostium of the coronary artery.

It would, however, be a great step forward in the understanding of coronaryphysiology, to be able to measure mean absolute coronary blood flow in selectivecoronary arteries and consequently absolute values of mean microvascular resistance.In chapter 5 a method based on continuous infusion thermodilution is proposedto determine absolute coronary blood flow. Besides the in-vitro study described inchapter 5, the method will be validated and applied in-vivo in both animals andhumans as well (chapter 6 and 7).

2.1.4 Treatment

When the presence of coronary artery disease has been confirmed, either by anatomicassessment or by physiologic measurements, the proper therapy should be applied.Usually the therapy is focussed on revascularisation of the coronary artery to restorecoronary blood flow. One possibility of revascularisation is percutaneous coronaryintervention (PCI). In PCI a guide wire is advanced into the coronary artery distal tothe location of the stenosis. Next, a balloon catheter, either with or without a stent, isadvanced over the guide wire until the location of the stenosis. The balloon is inflatedand the atherosclerotic plaque is pushed aside in order to expand the coronary lumen.A stent is used to prevent elastic recoil of the vessel wall.

It has been shown in clinical trials that stent placement results in a decreasedrestenosis rate after six months (20%) compared to balloon angioplasty alone(30%)(Erbel et al., 1998). However, in-stent restenosis might occur after stentplacement, largely caused by neointima formation. Drug eluting stents (DES) havebeen introduced with the prospect of reducing the restenosis rate. DES are coatedwith a polymer that slowly releases a drug which inhibits cell growth, and hence,neointima formation. However, only anatomic follow-up data related to these stentswere available so far, mostly obtained from intravascular ultrasound (IVUS) andquantitative coronary angiography (QCA) (Morice et al., 2002; Moses et al., 2003).As has been pointed out above, the functional severity of a (re-)stenosis does notnecessarily have to correspond to the anatomic severity of a stenosis. For that purposethe aim of chapter 3 was to perform hemodynamic measurements in DES stents bothimmediately after implantation and at follow-up to better understand the physiologicbehavior of these new type of stents and compare this to bare metal stents.

10 Chapter 2

Figure 2.3: The aortic valve. The aortic valve is located between the left ventricleand the aorta. Normally is consists of three cusps from two of which the coronaryarteries arise. As a result of fusion of the commissures turbulent flow through thevalve traumatizes the leaflets which eventually might lead to bacterial endocarditis.When the valve becomes severely obstructed, aortic valve replacement (AVR) mightbe performed. (Medical Illustrations copyright c©1999-2008 Nucleus Medical Art,Inc.All Rights Reserved. www.nucleusinc.com)

2.2 Valvular disease. The aortic valve

The aortic valve is situated between the left ventricle and the aorta. It consists of threecusps, formed by leaflets, and the aortic annulus. The cusps together with the sinusesof Valsalva form a space from which the coronary arteries emerge. There are threecusps: the right coronary cusp, from which the right coronary artery emerges, the leftcoronary cusp, from which the left coronary artery emerges, and the non-coronarycusp (figure 2.3).

In the case of an aortic stenosis an outflow obstruction is gradually developing.Depending on the etiology of the degeneration of the valve the outflow obstructionis the result of fusion of the commissures or stiffer properties of the leaflets due tocalcification and degeneration. The pathophysiological cause of the diseased valvecan be classified as acquired or congenital. From this classification acquired age-related degeneration of the aortic valve is the most common cause for aortic stenosis.Risk factors for degenerative aortic stenosis are often similar to those of vascular


Table 2.1: Severity of aortic valve stenoses by valve area and mean gradient.Severity AVA [cm2] Mean gradient [mmHg]Mild >1.5 <25Moderate 1-1.5 25-50Severere <1.0 >50AVA, aortic valve area

atherosclerosis namely diabetes, hypertension, hypercholesterolemia, and smoking,but might occur without such risk factors (Peltier et al., 2003).

Regardless of the etiology of the aortic valve stenosis, the left ventricle experiencesa pressure overload due to the valvular obstruction. Since the development of theaortic stenosis is a gradual process it allows adaptation of the left ventricle. Normalcardiac output is maintained by left ventricular hypertrophy, which may sustain largetrans-valvular pressure gradients for many years without reduction in left ventricularsystolic function or development of symptoms. Once symptoms occur, however,prognosis is poor if the aortic stenosis remains untreated. Angina pectoris, syncope,and dyspnea in exercise are the typical symptoms of a severe aortic stenosis.

2.2.1 Clinical assessment

Doppler-echocardiography has become the most important clinical technique todetermine the severity of an aortic stenosis. Moreover, it is the technique of choicefor follow-up of patients because of its non-invasive character. Besides the possibilityto distinguish bicuspid and tricuspid valves, trans-thoracic echocardiography is usedin detecting thickened and calcified leaflets as well (figure 2.4). When the aorticstenosis is detected, the left ventricular function as well as the degree of hypertrophycan be determined.

Doppler-echocardiography is essential in determining the aortic stenosis severity.Functional severity of the aortic stenosis is based on trans-valvular pressure gradientsand the aortic valve area. Using the Bernoulli equation in combination with Doppler-derived velocity measurements through the aortic valve and the left ventricularoutflow tract (LVOT), trans-valvular pressure gradients can be reliably estimated(figure 2.4).

The aortic valve area can be measured on an anatomic cross-sectional view of theaortic valve but is more reliably estimated using the continuity equation. To calculatethe aortic valve area the values for the blood flow velocities in the LVOT and throughthe aortic valve also as well as the dimensions of the LVOT should be determined.Classification of aortic stenosis severity is shown in table 2.1.

2.2.2 Treatment

Aortic valve replacement (AVR) is the treatment of choice in adults with an aorticstenosis. The decision to replace the aortic valve with an prosthetic valve is mostlysymptom driven. However, in asymptomatic patients with severe aortic stenosis or

12 Chapter 2

Figure 2.4: Example of echocardiography of an aortic stenosis. Top panel: Twodimensional parasternal long axis view of the heart. Left- and right ventricularcavity (LV and RV) as well as the left atrium (LA) and the aorta (Ao) can clearly bedistinguished. A clear bright echo at the location of the aortic valve (AoV) indicatesa calcified valve. Moreover, the interventricular septum (IVS) and the posterior leftventricular wall (PW) are thickened indicating that the aortic valve might cause anoutflow obstruction. Bottom panel: Doppler measurements of the blood flow velocitythrough the aortic valve. An estimate of a mean pressure gradient of 70 mmHg and apeak gradient of more than 100 mmHg confirms a severe aortic stenosis.


progressive left ventricular dysfunction, AVR is also indicated.Since the 1960s valvular prostheses have been commercially available. Nowadays

mono-leaflet tilting disk and bi-leaflet valves are most commonly used. The durabilityof most mechanical valves is excellent. The disadvantage of mechanical valves is theincreased risk of thromboembolic complications and consequently require long-termanticoagulation therapy. Biological valve prostheses on the other hand do not requirelong-term anticoagulation therapy but have limited durability.

Regardless of the type of prosthesis, the pressure gradient over the valve is reducedsignificantly after AVR. The reduced pressure gradient over the valve has a favorableeffect on coronary flow (Camici and Crea, 2007; Hildick-Smith and Shapiro, 2000).It has been suggested that the orientation of placement of mechanical valves alsoinfluences coronary blood flow (Bakhtiary et al., 2006; Kleine et al., 2002c). Inchapter 4 the influence of the orientation of mechanical valve prostheses on coronaryperfusion is investigated.

2.3 Peripheral vascular disease. The Abdominal AorticAneurysm

An abdominal aortic aneurysm (AAA) is a permanent dilatation of the abdominalaorta. By definition an abdominal aortic diameter >3 cm is considered to be anAAA. Nature and cause of the development of an AAA are unclear, however, smoking,hypertension, and age are strongly associated with aneurysms of the aorta.

The risk of developing an AAA is 4 to 9% in the male population over 65 yearsof age whereas the risk is 5 to 10 times as low in the female population of the sameage (Zipes et al., 2005, chap 53). The paramount concern of AAAs is the risk ofrupture. Eventually 1 in every 3 aneurysms will ultimately rupture (Darling et al.,1977) resulting in sudden death in 75% to 90% of the cases (Fleming et al., 2005).

2.3.1 Anatomic assessment. The diameter criterium

It has been shown that the risk of rupture strongly correlates with maximal diameterof an AAA (Glimaker et al., 1991). For AAAs smaller than 4.0 cm the risk of ruptureis 0.3% per year which increases to 1.5% per year for AAAs with a diameter between4.0 and 4.9 cm. Aneurysms of 5.0 to 5.9 cm have an annual risk of rupture of 6.5%(Brown and Powell, 1999). The risk of rupture for AAAs of 6.0 cm and larger increasessharply although this risk cannot be estimated accurately.

Because of the strong correlation with rupture, the maximal diameter is used inclinical practice for risk stratification. Generally a diameter of 5.5 cm and largeris considered an indication for surgery or endovascular repair. For clinical follow-up, ultrasonography is a practical and inexpensive method to determined maximalaneurysmal diameter. This method is, however, insufficient for planning operativerepair. Computed tomography (CT) or magnetic resonance imaging (MRI) aregenerally used for this purpose. Both techniques permit 3D reconstruction of the

14 Chapter 2

Figure 2.5: An example of an infra-renal abdominal aortic aneurysm (AAA). An AAAis a permanent dilatation of the abdominal aorta >3 cm. The normal abdominalaortic diameter ranges from 2 to 2.5 cm. (Medical Illustrations copyright c©1999-2008 Nucleus Medical Art,Inc. All Rights Reserved. www.nucleusinc.com)

entire aneurysm and have higher accuracy in determining the maximal diametercompared to ultrasonography.

2.3.2 Physiologic assessment. Hemodynamic measurements

Despite the small risk, aneurysms below accepted cut-off values do rupture and viceversa, many AAAs larger than the cut-off value of 5.5 cm will not rupture within thepatient’s life time (Conway et al., 2001; Powell and Brown, 2001). In an effort todevelop methods to better predict the risk of rupture, biomechanical parameters havebeen investigated.

Finite element wall stress analysis has shown to better correlate with rupture thanmaximal diameter alone (Fillinger et al., 2003). Such a stress analysis requires anaccurate geometric description of the aneurysm, usually obtained by CT or MRI.Additionally, wall properties are needed to perform patient specific stress analysis.Values for these mechanical properties are often based upon tensile test of excisedaneurysmal material.

Efforts have been made to determine values for wall properties in-vivo. Moreover,it has been shown that distensibility of the aneurysm was increased in patients whoexperience rupture (Wilson et al., 2003). To calculate the distensibility of an AAAboth the current hemodynamic load as well as the accompanying volume changeare required and should be determined simultaneously. In chapter 8 a study isperformed to estimate biomechanical properties of aneurysms larger than 5.5 cm bysimultaneously measured pressure and volume measurements.

Chapter 3

Hemodynamic evaluation ofcoronary stents

Coronary stenoses resulting from atherosclerosis, reduce maximal achievable coronary blood flowthrough the artery involved. Percutaneous coronary intervention using drug eluting stents or baremetal stents positively influences coronary physiology on the short term. Superior anatomic resultsare found for drug-eluting stents when compared to bare metal stents. Long term hemodynamiceffects, however, are not available for drug eluting stents in comparison to bare metal stents. In thisstudy we investigated physiologic parameters of coronary stents at implantation and at 6-monthfollow-up in the drug-eluting sirolimus stent and its bare metal counterpart implanted in pairswithin the same patient. Twenty patients, accepted for percutaneous coronary intervention (PCI)of at least two coronary arteries with comparable vessel- and stenosis characteristics, received atrandom one sirolimus-eluting stent and one bare metal stent. Hemodynamic measurements wereperformed just after stent implantation and at 6-month follow-up. The physiologic characteristicsof the drug-eluting sirolimus stents were superior to those of the equivalent bare metal stent.

Published in: M. van ’t Veer, N.H.J. Pijls, W. Aarnoudse, J.J. Koolen, and F.N. van de Vosse.Evaluation of the hemodynamic characteristics of drug-eluting stents at implantation and atfollow-up. European Heart Journal, 27:1811-1817, 2006.

15

16 Chapter 3

3.1 Introduction

Drug-eluting stents have been introduced with the prospect of reducing restenosis rateafter percutaneous coronary intervention (PCI). However, only anatomic data relatedto follow-up of these stents are available so far, mostly obtained from intravascularultrasound (IVUS) and quantitative coronary angiography (QCA) (Morice et al., 2002;Moses et al., 2003). It has been shown repeatedly that the physiologic parametersfractional flow reserve (FFR) and hyperemic trans-stent gradient (HTG) better reflectthe physiologic status of a coronary segment or stenosis both in native coronaryarteries and after stenting (Pijls et al., 1996, 2002b; Lopez-Palop et al., 2004).Moreover, it has been shown that there exists a relation between the wall shearstress (WSS) and the neointimal thickness (Wentzel et al., 2001; Gijsen et al., 2003).Therefore the aim of this study was to investigate FFR, HTG, and WSS at implantationand at 6-month follow-up in the sirolimus stent (CypherTM, Cordis, Johnson &Johnson, Miami, Florida) and in its bare metal counterpart (Bx VELOCITYTM),randomly implanted in pairs in two comparable arteries with comparable stenoseswithin the same patient.

3.2 Methods

3.2.1 Study population

Twenty consecutive patients with stable angina pectoris were selected. They wereaccepted for elective PCI of at least two coronary arteries with a comparablereference diameter and comparable stenosis characteristics. In none of the patientsprevious myocardial infarction had occurred in the myocardial regions supplied bythe respective arteries. The reference diameter of both arteries should vary less than0.5 mm and the length of the stenosis should not differ more than 50%. Patients withvery tortuous vessels, severe obstructive pulmonary disease, as well as patients witha contraindication for aspirin or clopidogrel were excluded. There were no furtherexclusion criteria.

All patients were selected from the population referred to our hospital for electivePCI of 2-vessel disease (figure 3.1). Among a total of 228 referred for elective 2-vesselPCI, 20 patients fulfilled the criteria of having comparable stenoses with comparablelength, severity, and reference diameter.

Immediately before the procedure, the placement of the sirolimus and thebare metal stent (BMS) in the two stenoses was determined by computer-codedrandomization. The study was approved by the institutional review board of theCatharina Hospital and written informed consent was obtained from all patients priorto the study. It should be noted that at the time the study was performed, in ourcountry, drug-eluting stents were not available for routine patients, except in studies.

Hemodynamic evaluation of coronary stents 17

Figure 3.1: Patient flow chart. Twenty consecutive patients fulfilling the studyinclusion criteria were selected from the total population of patients undergoing PCIreferred to our hospital in the period June 1st till December 3rd, 2003

18 Chapter 3

3.2.2 Interventional protocol

All procedures were performed by the femoral approach with 6F guiding catheters.Patients were pre-treated with aspirin and clopidogrel. Prior to the procedure 5000IU of heparin was administered. After intracoronary administration of 200 µgnitroglycerine, coronary angiography was performed. Next, a sensor-tipped pressureguide-wire (PressureWire 4, Radi Medical Systems, Uppsala, Sweden) was usedas routine guide wire in all of the procedures and pressure measurements wereperformed by this wire as described below. After successful stent implantation andrepeated measurement of coronary pressures, this pressure wire was replaced bya Doppler flow wire (FloWire, Jomed, Ulestraten, The Netherlands) for velocitymeasurements and WSS calculations as described below. Post-interventional phar-macologic treatment included aspirin and clopidogrel as routine in our center. Allinvasive measurements were repeated after 6 months with the same sequence andmethodology of all physiologic measurements.

3.2.3 Hemodynamic analysis

After adequate calibrating and positioning the pressure sensor in the distal part ofthe coronary artery, hyperemia was induced by intravenous infusion of adenosinethrough the femoral vein (140 µg/(kg ·min) ) as described before (Pijls et al., 1996;Pijls, 2004). After steady state maximum hyperemia had been achieved, FFR wasdetermined as the ratio of distal coronary pressure (Pd) and aortic pressure (Pa).

FFR expresses maximal achievable blood flow in the presence of a stenosis as aratio of normal maximal blood flow in the hypothetic case that the coronary arterywould be completely normal (Pijls et al., 1996; Pijls, 2004). Adenosine was stoppedand the stent was placed. Thereafter adenosine was started again and post-stentFFR was measured. Subsequently, during sustained hyperemia, the wire was slowlypulled back under fluoroscopic guidance and a pull-back recording was made (Pijlset al., 2002b; Pijls, 2004). Hyperemic trans-stent gradient (HTG) was calculated asthe pressure just proximal to the stent (Pprox) minus the pressure just distal to thestent (Pdist), both determined at maximum coronary hyperemia. Also the trans-stentpressure ratio (TPR) was calculated as the ratio of Pdist and Pproxduring maximumcoronary hyperemia. Next, adenosine was stopped and the pressure wire wasexchanged for a Doppler flow wire and average peak velocity (APV) was measuredat the entrance and the exit of the stent and within the body of the stent at rest.An approximation of WSS (τ) at the different positions was calculated assuming aPoiseuille flow yielding:

τ =8ηv

d(3.1)

where η is the dynamic viscosity of whole blood, v represents the averagecross-sectional velocity at the particular location, and d the corresponding diameterobtained from the QCA analysis (indicated in figure 3.2). Prior to the initialintervention and prior to the follow-up procedure, a venous blood sample was taken


Figure 3.2: Definition of the hemodynamic parameters in this coronary artery study.Pa, Pd, Pprox, and Pdist represent aortic pressure, distal pressure, pressure just proximaland just distal to the stent, respectively, all measured during maximal hyperemia. FFRrepresents the fractional flow reserve, HTG the hyperemic stent gradient, and TPRthe trans-stent pressure ratio. WSS indicates the approximate wall shear stress in thestent calculated using equation 3.1. The diameter of the stented vessel is indicatedwith ’d’.

before the administration of heparin for the determination of blood viscosity asdescribed elsewhere (Matrai et al., 1987). Because of the assumption of a parabolicPoiseuille flow in a circular straight tube the average velocity was taken half of themeasured APV (Buchi and Jenni, 1998). The definitions of the several hemodynamicindices are clarified in figure 3.2. Our method is fundamentally the same as usedby Wentzel et al. and Gijsen et al. However we do not use a finite element modelto calculate local wall shear stresses. Our equation is the analytic solution of theNavier-Stokes equations, under the assumptions of an incompressible steady laminarNewtonian flow through a straight tube.

3.2.4 Angiographic analysis

Angiograms were made after nitroglycerine administration in at least two orthogonalprojections and QCA was performed and analyzed as described before (Reiber et al.,1985). Reference diameter, percentage diameter stenosis, and minimal luminaldiameter (MLD) were calculated offline (QCA-CMS 4.0, MEDIS medical imagingsystems, Leiden, Netherlands) both before and immediately after the procedure, andat six-month follow-up period. Late lumen loss was defined as the difference betweenthe post-procedural MLD and the follow-up MLD at six months as described earlier(Morice et al., 2002; Moses et al., 2003). The diameters of the artery at the entranceof the stent, within the stent, and at the exit of the stent were measured in the twoprojections, averaged, and then used for the WSS calculations at those positions.

20 Chapter 3

3.2.5 Statistical analysis

The number of twenty patients was arbitrary but based on the consideration thatif a relevant difference between the stents would be present with respect to thehemodynamic indices, this should be demonstrable in this group. Because of theextensive invasive study protocol, it was considered inappropriate to include a largergroup of patients.

The design of the study introduces two sources of repeated measures that weconsidered for the analysis. First, measurements were performed immediately afterplacing the stents and repeated at follow-up. We calculated the difference betweenparameters for these time points and compared them across the stent groups. Second,two different stent types were compared within each patient. Therefore, we used theWilcoxon signed-rank test for paired observations. All tests were performed two-sided. A P-value of <0.05 was considered significant.

In performing statistical tests, no corrections were made with regard to the TypeI error as the primary outcome of this study was the difference between FFR fromimmediately after stenting to follow-up. Further tests may be considered as indicativefor the difference in the results between the stents. Statistical software package SAS(Version 8.2, SAS Institute, Cary, NC, USA) was used for the statistical analysis.

3.3 Results

3.3.1 Baseline characteristics and procedural results

All twenty patients eligible for the study consented to participate. Twenty sirolimusstents and twenty BMSs were implanted in pairs in twenty patients. Four patientsreceived an additional stent due to a residual significant pressure gradient elsewherein the vessel (n=3) or a dissection (n=1). These additional 4 stents were also drug-eluting stents. No procedural complications occurred. Patient characteristics arepresented in table 3.1. Baseline angiographic and hemodynamic characteristics aswell as stent characteristics are presented in table 3.2. No significant differenceswere present at baseline between the two groups.

3.3.2 Clinical follow-up at 6 months

There where no deaths among these twenty patients after six-month follow-up. Twocases of sub-acute stent-thrombosis occurred: one in a sirolimus and one in a baremetal stent. These events occurred within the same patient and were possiblyprovoked by removal of the sheath several hours after the intervention, accompaniedby bradycardia and a vasovagal reaction. Both stents were successfully re-opened byre-intervention. The CK-MB level rose to 677 U/l in this patient.

At six months, four re-interventions were necessary based upon an ischemic FFR:two cases of in-stent restenosis and the other two cases because of restenosis justproximal to the stent. All re-interventions were related to the BMS and were alltreated by placing a drug-eluting stent within or overlapping the former stent.


Table 3.1: Patient characteristicsPatients 20Male/female 16/4Age, years 59.1±10.5Risk factors

Smoking 9Diabetes 3Hypertension 9Family history 15Cholesterol, mmol/l 5.46±0.85HDL, mmol/l 1.28±0.29

MedicationAspirin 20Clopidogrel 19GP2B3A inhibitors 11Bta-blockers 16Statins 16ACE-inhibitors 6

HDL: High Density Lipids

Table 3.2: Baseline angiographic and pressure data before stent implantation andstent characteristics

Sirolimus BMS P-value(n=20) (n=20)

Angiographic parametersLesion length, mm 14.9±5.3 14.8±4.2 0.99Vessel RCA/Cx/LAD 8/5/7 8/5/7 -Lesion type (A/B1/B2/C) 6/5/8/1 7/5/6/2 -Reference diameter, mm 2.8±0.4 2.8±0.5 0.94MLD, mm 1.2±0.4 1.2±0.3 0.54Diameter stenosis, % 57±12 57±13 0.95

Pressure variableFFR 0.61±0.20 0.61±0.16 0.98

Stent characteristicsDiameter, mm 2.9±0.3 2.9±0.3 0.58Length, mm 17.8±5.7 16.5±3.7 0.46Inflation pressure, atm 14.7±1.5 14.6±1.7 0.81

RCA: Right coronary artery, Cx: Circumflex, LAD: Left anterior descending artery.

22 Chapter 3

3.3.3 Angiographic follow-up at 6 months

Immediately after intervention, the angiographic characteristics of both groups weresimilar. However, after six months they differed significantly with respect to the MLD,the percentage diameter stenosis, and the late lumen loss in favor of the sirolimusstent (table 3.3). The MLD of the sirolimus stent versus the BMS was 2.3±0.4 versus1.7±0.4 mm (p=0.041), the percentage diameter stenosis was 14±9 versus 36±15%(p=0.022), and the late luminal loss was 0.1±0.3 versus 0.6±0.5 mm (p=0.047),respectively.

3.3.4 Hemodynamic follow-up and WSS at 6 months

No significant differences were seen immediately after intervention between the twogroups with respect to the hemodynamic data and WSS (table 3.3). However, at six-month follow-up, both FFR and HTG of the sirolimus group differed significantly fromthe bare metal group. FFR was 0.91±0.05 versus 0.83±0.10 (p=0.028) and HTG was1.2±1.2 versus 7.5±8.1 mmHg (p=0.026) in favor of the sirolimus-stent. Also theTPR differed significantly: 0.99±0.01 in the sirolimus group versus 0.91±0.09 in thebare metal group (p=0.029) (table 3.3).

The normal reference value of WSS in a coronary artery at rest is 1.5-2 Pa (Ku,1997). There was no significant difference in wall shear stress for any of the positionsat baseline between the groups. The within stent values at six months differedsignificantly between the two groups (p=0.009). The values of WSS at the entranceand exit of the stent did not differ significantly at six months (table 3.3). Values forthe WSS at different positions are presented in figure 3.3.

3.4 Discussion

This study shows that drug-eluting sirolimus stents have a better and a more physio-logic hemodynamic performance at six-month follow-up than the corresponding baremetal stents. Moreover, at follow-up the sirolimus stent maintained normal values ofWSS in contrast to the bare metal stent, where high values of WSS were found withinthe stent.

It has been shown previously that calculating local WSS in stents is useful inidentifying specific locations at risk for restenosis (Wentzel et al., 2001; Gijsenet al., 2003). In those studies, WSS was calculated with a higher spatial resolutioncompared to our study where only global approximate values of WSS were obtained.Nevertheless our values are in the physiologic range of 1.5-2 Pa along the stentedsegment in both vessels just after stenting, the normal resting value for coronaryarteries in a wide range of species (Ku, 1997). In contrast to the BMS, the sirolimusstent maintained this normal WSS after six months. The high values of WSS found inthe BMS after six months reflect the decreased MLD and consequently the higherpercentage diameter-stenosis. These findings are in concordance with previousangiographic studies (Morice et al., 2002; Moses et al., 2003). As WSS also accounts


Table 3.3: QCA data and physiologic parameters immediately after stent implantationand at 6-month follow-up

Sirolimus BMS P-value(n=20) (n=20)

Angiographic parametersMLD, mmImmediately after PCI 2.4±0.3 2.3±0.4 0.236-month follow-up 2.3±0.4 1.7±0.4 0.023Change at follow-up -0.1±0.3 -0.6±0.5 0.041

Diameter stenosis,%Immediately after PCI 14±10 19±11 0.056-month follow-up 14±9 36±15 <0.001Change at follow-up -1±12 15±21 0.022

Late loss, mm 0.1±0.3 0.6±0.5 0.047Coronary pressure parameters

FFRImmediately after PCI 0.90±0.06 0.88±0.07 0.556-month follow-up 0.91±0.05 0.83±0.10 0.027Change at follow-up 0.01±0.05 -0.05±0.10 0.028

HTGImmediately after PCI 2.3±1.7 3.4±3.3 0.196-month follow-up 1.2±1.2 7.6±8.1 <0.001Change at follow-up -1.2±2.0 4.1±8.7 0.026

TPRImmediately after PCI 0.97±0.02 0.96±0.04 0.116-month follow-up 0.99±0.01 0.91±0.09 0.002Change at follow-up 0.01±0.02 -0.05±0.10 0.029

WSS (normal value: 1.5-2 Pa)Entrance of stentImmediately after PCI 2.0±1.2 2.0±1.0 0.936-month follow-up 1.8±0.8 2.4±1.4 0.24Change at follow-up -0.1±0.7 0.3±1.8 0.56

Within stentImmediately after PCI 1.9±0.8 2.0±1.0 0.436-month follow-up 1.6±0.7 3.9±3.1 0.003Change at follow-up -0.3±1.1 1.7±2.7 0.009

Exit of stentImmediately after PCI 1.9±0.8 2.1±0.8 0.46-month follow-up 1.7±0.9 1.9±1.1 0.34Change at follow-up -0.1±0.6 -0.2±0.8 0.7

MLD = minimal luminal diameter, PCI = percutaneous coronary intervention, FFR =fractional flow reserve, HTG = hyperemic trans-stent gradient, TPR = trans-stentpressure ratio, WSS = wall shear stress

24 Chapter 3

Figure 3.3: Approximate WSS values at three positions immediately afterimplantation and at 6-month follow-up for the sirolimus stent (top) and the baremetal stent (bottom). The locations marked with ’prox’ and ’dist’ indicated theposition just proximal and just distal to the stent, respectively.


for the average cross-sectional velocity and viscosity, it is a better indicator for thelocal hemodynamics within the stent than the anatomy-derived parameters alone.

Despite extensive studies on FFR and pressure gradients across bare metal stentsimmediately after implantation and at follow-up (Pijls et al., 2002b; Hanekampet al., 1999), little was known so far about those physiologic and hemodynamiccharacteristics of drug-eluting stents over time. The FFR and HTG we found in bothgroups show that after six months approximately half of the total pressure gradientpresent in the vessel stented with the BMS was due to gradient across the stent, incontrast to the vessel stented with the sirolimus stent where this gradient across thestent itself was very small (7.5±8.1 mmHg in the BMS versus 1.2±1.2 mmHg in thesirolimus stent respectively, p=0.026).

A recurring hyperemic gradient of 5-10 mmHg after six months in BMS, dueto intimal hyperplasia, has been described earlier (Pijls et al., 2002b; Hanekampet al., 1999). The present study shows that for sirolimus stents this phenomenonis much less severe. For both groups the pressure loss along the remaining non-stented part of the coronary artery was identical, indicating that the arteries in bothgroups were diseased to a similar degree with a diffuse hyperemic pressure decline ofapproximately 10 mmHg.

In patients with multiple but distant abnormalities within one coronary arteryand a significantly decreased FFR (<0,75-0.80), in the past it was not recommendedto stent spots or segments with a hyperemic gradient of <10 mmHg because, asmentioned above, despite optimal deployment a hyperemic gradient of 5-10 mmHgwas present again after six months in the majority of the bare metal stents (Hanekampet al., 1999). For sirolimus stents our study shows that the average HTG after 6months is significantly smaller, i.e. 1-2 mmHg. Therefore a practical implicationof this study for interventional cardiology is that in such diseased arteries withmultiple non-adjacent lesions, whether or not superimposed on diffuse disease orseparated by side branches and each in itself not hemodynamically significant but inseries responsible for inducible ischemia, the possibility for successful interventionaltreatment by several stents is significantly improved. The beneficial effect ofbare metal stents in those stenoses with gradient less than 10 mmHg was oftendisappointing in the past due to the recurring gradient of approximately 5-10 mmHgafter 6 months. Having established now that across DES only minor gradients arepresent at follow-up, interventional treatment in such patients has become morerationale. It should be emphasized that stenting on a purely anatomic basis in thesepatients makes little sense if FFR of all lesions together is >0.80 (Pijls et al., 2007).

Finally, because both the sirolimus stent and the BMS were implanted in pairsin stenoses with comparable characteristics within the same patient, the biologicalenvironment and risk factors were identical as much as possible and the possibilitiesthat the differences observed were due to other factors than the stent itself wasminimized as much as possible.

26 Chapter 3

3.4.1 Limitations

Calculation of WSS in this study was based upon QCA, APV, and viscosity, limitedto the entrance, exit, and body of the stent. These calculations are influenced bythe inaccuracy in APV which has been described before (Buchi and Jenni, 1998). Tohave obtained accurate geometrical information and consequently accurate numericalshear stress data, IVUS would have been necessary. However, because this wouldhave further prolonged the time of these extensive procedures, we chose for WSScalculations by equation 3.1 at the entrance, body, and exit of the stent only, asexplicated in section 3.2.3.

Although we did not specifically investigated inter and intra observer variability,we assumed that this would be limited, because WSS is calculated directly fromviscosity, flow velocity, and QCA, all of which have a small inter and intra observervariability (Matrai et al., 1987; Buchi and Jenni, 1998; Reiber et al., 1985). Withrespect to the limited number of stenoses (2 × 20), it should be noted that itwas not the intention of this study to demonstrate any difference in restenosis- oradverse event rate, but to acquire better understanding of the physiologic behaviorof sirolimus stents compared to the BMS, which was clearly achieved in this study(table 3.3).

3.5 Conclusions

At six-month follow-up the sirolimus stent was superior compared to its baremetal counterpart not only with respect to angiographic but also to physiologiccharacteristics. Fractional flow reserve was significantly higher and the hyperemictrans-stent gradient significantly lower for the sirolimus stent. Furthermore, incontrast to the bare metal stent, the sirolimus stent maintained a normal wall shearstress within the stented segment.

Acknowledgements

This study was supported by a grant of Cordis, a Johnson & Johnson Company andby grant 04-03 of the foundation ’Stichting Vrienden van het Hart’, Eindhoven theNetherlands. The authors are indebted to the nursing staff of the catheterisationlaboratory of the Catharina Hospital in Eindhoven for their dedicated assistance inthe invasive procedures.


Editorial: Hemodynamic findings after drug-elutingstenting: expected, provocative, or challenging?

This editorial by Fernando Alfonso refers to ’Evaluation of the hemodynamic characteris-tics of drug-eluting stents at implantation and at follow-up’ (van’t Veer et al., 2006)

Published in: European Heart Journal 2006;27:1764-1766, printed withpermission of the European Heart Journal, Oxford University Press

In this article van’t Veer et al. present a comprehensive and detailed studycomparing the hemodynamic characteristics of drug-eluting stents (DES) with thoseobtained by conventional bare metal stents (BMS). After DES, long-term physiologicparameters including fractional flow reserve (FFR), hyperaemic gradient, and wallshear stress (WSS) were superior to those found in equivalent BMS implanted inthe same patients (van’t Veer et al., 2006). Although these findings are of majorinterest, most of the new information provided could be perceived as well expectedconsidering the large body of evidence demonstrating the superb late results afterDES implantation. Nevertheless, as will be highlighted in this editorial, some studyfindings and their implications are rather provocative. Furthermore, on the basis oftheir results, these investigators from the Catharina Hospital (Eindhoven)(van’t Veeret al., 2006) dare to challenge some widely accepted strategies in the management ofpatients with diffuse coronary artery disease (CAD).

Coronary physiology after DES

Previous land-mark studies from the same group have unequivocally established thesuperiority of FFR over conventional angiography to assess the functional severityof coronary stenosis (Pijls et al., 1996). Even intravascular ultrasound (IVUS),able to provide a thorough anatomic coronary assessment, can only be used as asurrogate of lesion physiology (Alfonso et al., 2003). Large-scale serial morphologicalstudies have consistently demonstrated the unique ability of DES to prevent restenosisand to inhibit neointimal proliferation. However, functional studies after DESimplantation using direct hemodynamic assessment are scarce (Gijsen et al., 2003;Carter et al., 2005). The elegant study of van’t Veer et al. assessing FFR, stent-inducedgradients, Doppler-derived intracoronary velocities, and WSS, fills the gap in ourunderstanding of DES influence on coronary physiology. The study design, randomlyallocating in pairs DES and BMS in well-selected matched lesions of patients withtwo-vessel disease, circumvents the potential confounding influence of systemic andanatomic factors on outcome measures. This sound methodology enables to obtainmeaningful hemodynamic information from a relatively small patient cohort, which,in turn, is critical when relatively sophisticated diagnostic procedures are performedduring coronary interventions (Alfonso et al., 2003). Likewise, excluding unstablelesions, infarct-related vessels, and selecting the intravenous approach to administeradenosine minimizes potential pitfalls in physiologic measurements.

28 Chapter 3

A previous study demonstrated that FFR immediately after BMS was able topredict long-term clinical outcome (Pijls et al., 2002b). Intuitively, it is difficult toanticipate a similar application for this parameter after DES, considering their lowrestenosis rate. Likewise, DES thrombosis is exceedingly rare (although occurred inone patient of the present series).

Therefore, the routine assessment of hemodynamic factors is not justified.However, from an academic stand point, the excellent long-term hemodynamicfindings obtained after DES are re-assuring.

WSS in atherogenesis and neointimal proliferation

Atherogenesis

WSS is the tangential drag (frictional) force produced by flowing blood on theendothelial surface. WSS is probably the most important local factor influencingatherogenesis (Stone et al., 2003; Irace et al., 2004). It has been suggestedthat atherosclerosis predominantly develops at segments with low WSS. Low WSSinduces endothelial dysfunction, inflammation, and smooth muscle cell proliferation.Atherosclerosis frequently has an eccentric distribution and preferentially occurs inthe proximal coronary segments, at bifurcations, and in the inner curve of theartery. Typically, all these locations have low WSS. In particular, at the hips of thebifurcation (walls opposite to the flow divider), low WSS values are systematicallydetected. According to the HagenPoiseuilles law, WSS is inversely proportional tothe cube of the radius explaining its dramatic relation with the lumen size. Althoughsome investigators have hypothesized that a threshold level of WSS is required toaffect atherogenesis, the boundary between atheroprotective and atherogenic effectsremains as yet undefined (Irace et al., 2004). Moreover, the contribution of WSS toatherogenesis appears clear in low-risk individuals, but its effects might be masked inhigh-risk subjects or in mature lesions (Irace et al., 2004). For precise local WSScalculations, the non-linear, incompressible fluid, three-dimensional NavierStokesequations (governing the conservation of mass, energy, and momentum) need to besolved. This form of detailed virtual analysis (computational fluid dynamics) is nowreasonably practical using specialized programs for the discretization of flow usingfinite-element methods (Stone et al., 2003; Irace et al., 2004; Wentzel et al., 2001;Carlier et al., 2003; Sanmartin et al., 2006). A mesh is generated and adequateboundary conditions are defined. Local WSS is a sophisticated parameter that maybe calculated only after a comprehensive anatomical and physiologic assessment.This requires a true three-dimensional anatomic reconstruction of the vessel inrelation to the distribution of intravascular velocity profiles. In previous studies,biplane angiography combined with IVUS was used for accurate volumetric lumenreconstruction (Stone et al., 2003; Irace et al., 2004; Wentzel et al., 2001; Carlieret al., 2003; Sanmartin et al., 2006). Alternatively, WSS may also be measured using aglobal analysis at different coronary segments, as in the study of van’t Veer et al.. Thisprovides a valid approximation to WSS at selected vessel positions highly attractivein the clinical setting.


Neointimal proliferation

Neointimal thickness distribution after BMS is related to local factors including extentof vascular injury, inflammation, and WSS. Neointimal tissue tends to proliferate inareas with low WSS, whereas the stent remains free of cell proliferation in areaswith high WSS (Wentzel et al., 2001; Carlier et al., 2003). Of interest, stent designand deployment techniques influence WSS. Experimentally, increasing WSS using aflow divider within BMS, has been accompanied by local reductions in neointimalhyperplasia, inflammation, and wall injury (Carlier et al., 2003). Although evidencesuggests that WSS influences neointima proliferation after BMS, its exact role remainscontroversial (Stone et al., 2003).

Preliminary data of WSS after DES are particularly intriguing. In the study ofGijsen et al. biplane angiography and IVUS were used to determine volumetric lumengeometry 6 months after DES. Flow velocities were directly recorded within the stent,and WSS was obtained from computational fluid dynamics. A significant inverserelation was found between WSS and DES neointimal thickness. More recently, Carteret al. conducted serial analysis of segmental WSS after oversized BMS and DESimplantation in a porcine model. Relatively low WSS was induced after deploymentwith both stents. However, at 30 days, IVUS-derived lumen areas were larger andnormalized WSS was lower after DES. A negative correlation was found betweenWSS immediately after BMS and the subsequent neointimal formation. Unexpectedly,post-DES WSS had a positive correlation with the neointimal proliferation. The studyof van’t Veer et al. demonstrated that WSS at follow-up is higher in patients treatedwith BMS than in those treated with DES. Although such findings might be expectedbecause of the poorer angiographic results of the former group, this concept deservesfurther attention. In particular, the ability of DES to inhibit neointimal proliferationcritically depends on their initial pharmacologic action. Later on, when the drug effecthas vanished, an excellent hemodynamic profile, with physiologic WSS patterns, mayfurther prevent cell proliferation. However, in this study, the mean in-stent WSSafter DES was 1.9±0.8 Pa but only 1.6±0.7 Pa at follow-up. Therefore, at leastin some patients, relatively low late WSS values were found. Whether low WSS 6months after DES could promote a delayed neointimal response remains speculative.Conversely, one may also suggest that a negative feedback control loop may occurafter BMS. In this scenario, cell proliferation and the resulting lumen narrowingsignificantly increase WSS which, at least on theoretical grounds, might preventfurther neointimal growth. The potential contribution of high WSS to reduce theextent of late neointimal obstruction is also largely speculative. In the present study,however, the potential long-term implications of the hemodynmic parameters seenimmediately after stenting were not analysed.

To the best of our knowledge, the attractive methodology used in this study hasnot been previously validated in human coronary arteries; therefore, reproducibilitydata might have been of interest. In contrast, the correlation of these measurementswith those found using the classical approach to determine local WSS warrantsfurther studies. Finally, most WSS analyses currently neglect flow pulsatility, coronarymotion, and wall distensibility. Eventually, the challenge remains to develop a robust

30 Chapter 3

and well-validated tool to readily assess the value WSS in the clinical setting.

Implications for management of diffuse CAD

Some of the authors conclusions in this regard are rather provocative. The virtualabsence of measurable gradients along the entire stent length at late follow-upopens the Pandoras box of the treatment of diffuse CAD. Currently, interventions ondiffuse disease are only advocated when angiographically significant narrowings areidentified, ideally associated with evidence of ischaemia. Otherwise, angiographicallymild lesions on diffusely diseased long segments are usually left untreated despitetheir ability to generate gradients under maximum hyperaemia (de Bruyne et al.,2001a). Considering that significant gradients may be observed at follow-up afterBMS implantation (despite persistence of good angiographic results), the full-metal-jacket approach does not provide an attractive solution for this elusive form ofangiographically mild, but hemodynamically significant, disease. Thus, BMS spotstenting after angiographic, IVUS, or FFR guidance has been proposed in this vexingscenario. Current data from the Eindhoven group (van’t Veer et al., 2006) suggestthat treatment of diffuse coronary segments with long or multiple DES might solvethis situation. The absence of significant gradients at follow-up suggests that thisapproach could be used in selected patients. However, we should keep in mind thatthe risk of thrombosis after DES implantation is related to stent length. Accordingly,before this critical step is contemplated, more information is eagerly required. Onlyadequately designed trials, with well-established clinical and angiographic endpoints,will help to elucidate the potential clinical benefit of this provocative and challengingstrategy in selected patients with diffuse disease.

Chapter 4

Influence of orientation ofmechanical bi-leaflet valve

prosthesis on coronaryperfusion pressure

Orientation of a mechanical bi-leaflet valve prosthesis in aortic position might influence coronaryperfusion. The aim of this study was to investigate the influence of the orientation of amechanical bi-leaflet valve prosthesis on coronary perfusion pressure during hyperemia andadrenergic stimulation. During hyperemia perfusion pressure determines coronary blood flow.Fourteen patients with normal coronary angiogram underwent aortic valve replacement by abi-leaflet prosthesis, and 7 received a bio-prosthesis. Patients receiving a bi-leaflet prosthesiswere randomized to either orientation A (hinge mechanism perpendicular to a line drawnbetween the coronary ostia) or B (hinge mechanism parallel to the line between the ostia).Six months after surgery all patients underwent cardiac catheterisation. Coronary perfusionpressures were measured during resting conditions, during maximum hyperemia, and duringmaximum adrenergic stimulation with a guiding catheter in the aortic arch, simultaneouslywith a sensor tipped guide wire in the coronary artery and in the aortic root. Only small non-significant differences in coronary perfusion pressure were found between different orientations ofa mechanical bi-leaflet prosthesis or between bi-leaflet prostheses and bio-prostheses.

Published in part in: M. van ’t Veer, A.H.M van Straten, F.N. van de Vosse, N.H.J. Pijls.Influence of orientation of bi-leaflet valve prostheses on coronary perfusion pressure in humans.Interactive Cardiovascular and Thoracic Surgery, 6:588-592, 2007.

31

32 Chapter 4

4.1 Introduction

A well known effect of aortic valve replacement (AVR) in patients with aortic valvestenosis is an improved coronary flow reserve (CFR) after surgery (Camici and Crea,2007; Hildick-Smith and Shapiro, 2000). After AVR the pressure gradient over thevalve decreases significantly which has a favorable effect on the intraventricularcompressive forces and consequently on the coronary perfusion pressure as comparedto left ventricular cavity pressure (Nemes et al., 2002). The increased CFR ismainly attributed to a lower coronary resting flow. Recently is has been suggestedthat coronary blood flow is affected also by different types and orientations ofaortic valve prostheses (Bakhtiary et al., 2006; Kleine et al., 2002c). Kleine et al.performed animal experiments during which peri-operatively the orientation of thevalve prosthesis could be changed and absolute coronary flow was recorded. Changesin coronary blood flow were found as response to different orientations of the valves.These changes were ascribed to a disturbed flow pattern in the proximal part of theaorta downstream the valve originating in systole and having effect on coronary blooduntil diastole (Kleine et al., 2002b).

When blood is forced through the orifice of the valve prosthesis, the bloodflow accelerates thereby increasing the kinetic energy locally at the cost of thehydraulic energy, the pressure. Hypothetically, maximum velocity and thereforeminimum pressure (venturi effect) is observed just distal to the prosthetic valve atthe position of the coronary ostia as was shown by Fiore et al. (2002). Consequently,differences in coronary perfusion pressure caused by a venturi-like effect, in differentorientations of a mechanical bi-leaflet prosthesis (BLP) during hyperemia andadrenergic stimulation, could influence coronary blood flow.

It should be noted that coronary flow is regulated by a complex interactionof neural and humoral responses, depending on oxygen demand and affectingmyocardial resistance. The resistance can be minimized (and therefore kept constant)by pharmacologic stimuli like adenosine and dobutamine, reflecting maximumexercise in true life. Under such circumstances coronary blood flow has been shownto be directly proportional to coronary perfusion pressure (Pijls, 2004; Pijls et al.,1996, 1993). Under resting circumstances and in the absence of significant coronaryartery stenosis, any drop in coronary perfusion pressure is compensated by a decreaseof coronary arteriolar resistance (autoregulation). But under maximum adrenergicconditions, reflecting exercise, arterial resistance reserve is exhausted on one hand,whereas the transvalvular gradient is largest. Therefore under such circumstancesdifferences in perfusion pressure influence coronary blood flow.

The main goal of the current study was to investigate if there were clinicallyrelevant differences in coronary perfusion pressure for different valve prostheses anddifferent orientations of a mechanical bi-leaflet prosthesis (BLP) in conscious humansduring pharmacologic-induced hyperemia and dobutamine-induced adrenergic stim-ulation, reflecting exercise in true life.

Since absolute flow measurement during cardiac catheterisation is not possible sofar, surrogate measures for coronary blood flow were used to interrogate coronaryphysiology besides pressure measurements, i.e. coronary blood flow velocity mea-

Valve prostheses and coronary perfusion 33

Figure 4.1: Schematic representation of the orientations as defined for the bi-leafletmechanical valve prosthesis. Hinge mechanisms placed perpendicular to a line drawnbetween the coronary ostia for orientation A (left) and parallel for orientation B(right). LCA and RCA represent the left and the right coronary artery respectively.

surements (White, 1993). Coronary blood flow velocity was determined by placing aguidewire with a Doppler crystal tip in the coronary artery. Direction and magnitudeof the flow velocity could be measured in this way. As a secondary goal of the currentstudy, epicardial flow reversal was investigated. It might be expected that in the caseof a significant venturi effect, perfusion pressure drops, and epicardial flow reversalcan be observed.

4.2 Materials and Methods


A total of 21 patients accepted for isolated AVR and a normal coronary angiogramwas selected from the total population referred to our hospital. Exclusion criteriawere: previous valve surgery in any position; concomitant surgical procedures ascoronary artery bypass grafting, mitral or tricuspid valve repair or annuloplasty, orresection of a left ventricular aneurysm; evidence of left ventricle dilatation; aorticregurgitation ≥ grade 2; history of myocardial infarction; pre-operative creatine level> 150 µmol/l; or age over 75 years at time of surgery. Fourteen patients wereselected to receive a mechanical BLP and seven patients received a bio-prosthesis.Prior to surgery patients receiving a mechanical BLP were randomized to differentorientations of the valve. We defined two orientations: A and B, comparable toother studies (Kleine et al., 2002c,b). In orientation A (orthotopic position) the hingemechanism was placed perpendicular to a line drawn between the coronary ostia, inorientation B (heterotopic position) the hinge mechanism was placed parallel to aline drawn between the two coronary ostia (figure 4.1).

34 Chapter 4

4.2.2 Echocardiography

Prior to surgery and at six month follow-up echocardiography was performed.Dimensions of the left ventricle were measured: thickness of the interventricularseptum (IVS), the inner diameter of the left ventricle (LVID), and thickness of theposterior wall (LVPW). These measurements were taken in diastole at the peak ofthe R-wave in a parasternal long-axis view. Trans-valvular pressure gradients wereestimated from blood flow velocities measured by Doppler echocardiography in theleft ventricular outflow tract and in the ascending aorta assuming inviscid flow.

4.2.3 Surgical procedure

Anaesthesia was performed according to routine in our hospital. All patientsunderwent a sternotomy, followed by arterial cannulation in the ascending aorta andsingle venous cannulation in the right atrium. Operations were performed undernormothermic extra-corporeal circulation and the heart was arrested by the use ofintermittent warm blood or cold crystalloid cardioplegia by antegrade administrationaccording to the surgeons preference. The native aortic valve was removed andthe annulus was decalcified. The valve prosthesis was implanted using running orseparate sutures again according to the surgeons preference. Mechanical prostheseswere implanted with an orientation according to randomization. Most patients wereextubated within a few hours and left the ICU the same day.

4.2.4 Catheterisation at follow-up

Follow-up catheterisation was performed six months after surgery. Forty-eight hoursprior to the procedure, patients were asked to stop beta-blockers in order to avoidinhibition of the maximal adrenergic stimulation by dobutamine. catheterisation wasperformed by the Judkins approach using 6F guiding catheters with side holes ata distance of 10 cm from the tip, enabling pressure recording at the level of theaortic arch (Zuma, Medtronic Inc., Minneapolis, MN, USA). After administration of5000 IU of heparin, the guiding catheter was advanced into the ostium of the leftcoronary artery, 200 µg of nitroglycerine was administered, and control angiogramswere made to visualize the coronary artery. Next, a 0.014” pressure guide wire(PressureWire 5, Radi Medical Systems, Uppsala, Sweden) was advanced into theproximal coronary artery and pressure measurements were performed as describedbelow under resting circumstances, during maximum coronary hyperemia by intra-venous adenosine administration, and during maximum adrenergic stimulation byintra-venous administration of dobutamine. Finally the pressure wire was replacedby a Doppler flow wire (FloWire Jomed, Ulestraten, The Netherlands) for coronaryblood flow velocity measurements. At the end of the procedure, 5 mg of metroprololwas administered to counteract the effect of dobutamine.


Pressure measurements

Pressure Pao was measured in the aortic arch by the fluid filled guiding catheter atthe location of the side holes. All other pressures were compared to this value asa gold standard for arterial pressure. Pressure measured by the pressure wire wasdetermined at three locations: the location of these side holes called Pao,pw; at thelevel of the ostium of the coronary artery (considered aortic root pressure) calledProot; and pressure in the proximal coronary artery called Pcor (figure 4.2).

At first, pressure measurements were performed under resting conditions. Thepressure wire was advanced until the pressure-sensor was at the location of the sideholes. At that position equilibration with the fluid-filled catheter was performed(Pao= Pao,pw). Thereafter, the two pressure signals were recorded simultaneouslythroughout the procedure. The wire was advanced into the proximal part of thecoronary artery to record Prootand further advanced to record Pcor. Under normalcircumstances all these pressures should be equal (de Bruyne et al., 1996).

Next, steady state maximum coronary hyperemia was induced by intra venousinfusion of adenosine in a dosage of 140 µg/(kg ·min) as described before (Pijls et al.,1996). During hyperemia, the pressure wire was slowly pulled back into the aorticroot. In case of a coronary artery ostial stenosis, whether iatrogenic by the precedingAVR or atherosclerotic, a gradient is expected between Pcor and Proot(Pijls, 2004).Next, the sensor was pulled back to the level of the side holes of the guiding catheterto ensure equal pressures again. After the adenosine administration was stopped,its effect vanishes within a couple of minutes. Thereafter, the pressure wire wasadvanced into the coronary artery again and dobutamine i.v. was started at a rate of10 µg/(kg ·min) and stepwise increased with 10 µg/(kg ·min) every two minutes untilthe maximum dosage of 40 µg/(kg ·min) was reached. It has been shown previouslythat dobutamine in this dosage not only stimulates contractility and ejection velocityof blood across the aortic valve, but also induces maximum hyperemia (Bartuneket al., 1999).

At maximum adrenergic stimulation, the pressure guide wire was pulled backagain in a similar way as during maximum hyperemia by adenosine. The differencebetween the root and the arch (Pao- Proot) indicates presence of a venturi pressuredrop due to artificial valve induced high ejection velocities across the valve.

Coronary blood flow velocity measurements

While continuing the dobutamine infusion at the highest dosage, the pressure wirewas removed and replaced by a FloWire. The FloWire was advanced into the proximalpart of the coronary artery to study phasic blood flow patterns during systole anddiastole and to investigate if and to what degree retrograde systolic blood flow waspresent. This phenomenon of retrograde flow has been described earlier in patientsafter AVR during resting conditions (Fujiwara et al., 1989; Kenny et al., 1994) and inpatients with aortic valve stenosis during adrenergic stimulation (Petropoulakis et al.,1995).

36 Chapter 4

Figure 4.2: Schematic representation of the pressure measurements in the aorta. Thegray fluid filled catheter is used as a reference pressure and measures pressure at theside holes (Pao). The pressure wire is advanced or pulled back through the guidingcatheter and measures pressure at the aortic arch (Pao,pw), the aortic root (Proot), andthe proximal coronary artery (Pcor). LCA and RCA represent the left and the rightcoronary artery, respectively.

4.2.5 Data analysis and statistical analysis

When studying pressure gradients across the coronary ostium it should be kept inmind that this pressure gradient is proportional to aortic pressure during maximalcoronary hyperemia. Therefore, gradients across the ostium (Proot- Pcor) werenormalized by taking changes in aortic pressure into account. The comparison of thegroups was performed by the non-parametric Mann Whitney test. Data are expressedas mean±SD or as median with interquartile range (IQR). A p-value of 0.05 wasconsidered statistically significant.

4.3 Results

4.3.1 Procedural results

Out of our 21 patients, seven patients had a BLP in position A (BLP A), sevenin position B (BLP B), and seven had a bio-prosthesis (Bio). The pre-operativebaseline characteristics of these patients are given in table 4.1. Average diametersof the prostheses for the different groups were 23.6 mm, 22.9 mm, and 23.9 mm,respectively for the BLP A, BLP B, and Bio.

Two patients refused angiographic follow-up (1 BLP B and 1 Bio). The six monthsfollow-up catheterisation with hemodynamic measurements could be performedwithout problems in all remaining 19 patients. No adverse events or complicationsoccurred. In all patients, the protocol could be followed as described above for theleft coronary artery.


Table 4.1: Preoperative patient characteristicsPatients 21Male/female 13/8Age,years 63.7±8.4

Risk factorsDiabetes 3Systemic hypertension 10Smoking 5

Cardiac rhythmSinus 19Paced 1Atrial fibrillation 1Peak gradient, mmHg 90±21AVA, cm2 0.87±0.25

MedicationAspirin 8Clopidogrel 1Acenocoumarol 1Bta-blockers 9Statins 8ACE-inhibitors 7Diuretics 10

Values are mean±SD; AVA, aortic valve area

Table 4.2: Echocardiographic data before surgery and during follow-upPre FU

BLP A BLP B Bio BLP A BLP B BioMean Ao gradient [mmHg] 57(24) 54(24) 51(14) 11(3) 16(13) 12(9)Peak Ao gradient [mmHg] 88(31) 84(27) 85(39) 21(6) 29(20) 28(13)IVS [mm] 16(7) 12(1) 19(8) 12(4) 12(2) 14(4)LVID [mm] 48(9) 46(10) 44(11) 48(12) 47(11) 40(5)LVPW [mm] 14(8) 11(2) 15(6) 11(5) 11(2) 12(0)

Values are median(IQR);BLP A and B,bi-leaflet prosthesis orientation A and B; Bio,bioprosthesis; Ao, aortic.

38 Chapter 4

Table 4.3: Physiologic data during the measurementsBLP A BLP B Bio All

Pao HR Pao HR Pao HR Pao HRRest 96(25) 81(26) 99(19) 79(23) 100(45) 78(17) 96(23) 81(19)Adeno 89(29) 85(28) 89(18) 88(23) 93(34) 93(21) 89(25) 88(24)10 dobu 96(9) 70(21) 95(15) 78(27) 110(14) 78(11) 99(17) 75(19)20 dobu 100(7) 89(64) 94(21) 108(24) 104(38) 95(27) 100(21) 104(34)30 dobu 92(17) 105(50) 85(24) 123(24) 87(14) 107(27) 92(16) 117(34)40 dobu 96(24) 125(39) 82(27) 131(22) 81(16) 119(3) 91(25) 128(25)

Values are median(IQR); Pao, aortic pressure in mmHg; BLP A and B,bi-leaflet prosthesis orientation A and B; Bio,bioprosthesis; HR, heart rate; adeno and dobu, adenosine and dobutamine administration intra-venously.

4.3.2 Echocardiographic results

Echocardiographic results before surgery and at six months follow-up are shown intable 4.2. Except for the expected significant reduction in pressure gradients over theaortic valve, no significant differences were present for any of the echocardiographicparameters. No differences were found at follow-up between orientations A and B ofthe mechanical valves or between the biological and mechanical prostheses.

4.3.3 Hemodynamic measurements

Pressure measurements

The values of blood pressure and heart frequency during the different steps of theprocedure for the two orientations and types of aortic valve prostheses are shownin table 4.3. A slight overall decrease in mean aortic pressure was observed for allpatients with increasing dosages of dobutamine (96(23) mmHg to 87(25) mmHg,p=0.021). The heart frequency on the other hand increased from 81(19) to 128(25)beats per minute, (p=0.011) as expected. The results of the coronary perfusionpressure measurements are shown in figure 4.3. Panel A and B describe the pressuregradients (Proot- Pcor) in resting conditions and during adenosine induced hyperemia.No severe proximal coronary or ostial disease was present in any of the groups. PanelsC and D express the venturi-related pressure drop (Pao- Proot) under resting conditionsand during sympathetic stimulation by dobutamine infusion. Although a smallpressure drop is present at the level of the coronary ostium, no significant differencebetween the different valve orientations or between the different prostheses wasobserved.

Coronary blood flow velocity measurements

Phasic coronary blood flow velocity data were assessed qualitatively to investigatesystolic retrograde flow during maximal adrenergic stimulation by dobutamine. Forthe BLP A we observed retrograde flow in 3 out of 7 patients, in the BLP B group in 3out of 6 patients, and in the bio-prosthesis group in 4 out of 6 patients. An exampleof retrograde flow during maximal adrenergic stimulation is presented for a patientfrom the BLP A group in figure 4.4 panel B.


Figure 4.3: Box-and-whisker plots for the pressure gradients representing ostialabnormalities ((Proot- Pcor); iatrogenic or atherosclerotic) during resting conditionsand during adenosine induced hyperemia (panel A and B). Box-and-whisker plotsfor the pressure gradients representing the velocity induced venturi effect (Pao- Proot)during resting conditions and during dobutamine induced adrenergic stimulus (panelC and D). BLP A and BLP B represent bi-leaflet prosthesis in orientation A and B; Biorepresents bio-prosthesis.

40 Chapter 4

4.4 Discussion

Coronary perfusion pressure

In this pilot-study in 21 patients undergoing aortic valve replacement, only anegligible influence of different orientations of the BLP on coronary perfusionpressure was found. Moreover, coronary perfusion pressure was not different betweenmechanical and biological prostheses.

In our protocol we differentiated between resting conditions, maximal pharmaco-logic coronary hyperemia, and maximal adrenergic stimulation in combination withmaximal coronary hyperemia (to mimic maximum exercise in true life). Therefore,we measured pressures in the coronary artery and in the aortic root simultaneouslywith the pressure in the aortic arch. We are not aware of another way to measurecoronary perfusion pressures more accurately. Recently it has been suggested inacute animal experiments that coronary blood flow is affected by the orientation ofa mechanical valve substitute with respect to the origin of the coronary arteries. Inthese experiments higher coronary flow rates were observed in one orientation ofmechanical valves (Kleine et al., 2002c). This specific orientation, comparable withthe orthotopic orientation A in our study, was defined as ’optimal’ in previous studiesof Kleine et al. with respect to hemodynamic parameters measured peri-operatively(Kleine et al., 2002b, 2000, 2002a). It was hypothesized from these studies thatdifferences in aortic root turbulence could explain the differences in coronary bloodflow.

Coronary blood flow is regulated, depending on oxygen demand and by a complexinteraction of neural and humoral responses affecting myocardial resistance. In theperi-operative setting this regulatory mechanism is disturbed, among others by theadministration of cardioplegia which induces a decrease in coronary artery resistanceand blunts autoregulation.

As we performed our measurements 6 months after successful AVR and notonly under resting conditions but also during maximum hyperemia and adrenergicstimulation, we believe that our results are representative for the physiologic situationafter valve surgery. Autoregulation has been restored at that time and the testingprotocol was representative for maximum exercise in true life. We did not find anydifferences in perfusion pressure for the two orientations of the BLP neither for thebio-prostheses. Additional information on proximal or ostial coronary disease wasobtained by the pressure sensor pullback recording during adenosine infusion (Pijls,2004).

Assuming that a venturi-induced pressure drop could influence coronary bloodflow, it is expected that these effects would be most pronounced during high cardiacoutput, i.e. during adrenergic stimulation. During our experiments we administereddobutamine in a high dosage simulating exercise by stimulating contractility andejection velocity of blood across the aortic valve. Next, a pressure sensor pullbackrecording was performed only slight increases in pressure gradients were foundbetween the aortic arch and the aortic root (Pao- Proot) compared to resting conditions.These pressure gradients were so small that any influence on coronary perfusion


Figure 4.4: Example of a blood flow velocity measurement during resting conditions(panel A) and during maximal adrenergic stimulation with dobutamine (panel B) ina patient with a bi-leaflet prosthesis in orientation A. At rest, an anterograde systolicblood flow is observed whereas at maximal adrenergic stimulation total flow stronglyincreases but reversal of blood flow occurs in early systole.

pressure and therefore coronary blood flow was clinically irrelevant. In this respect,it should be realized that in normal hearts, maximum perfusion should decrease by atleast 25% to result in inducible ischemia (Pijls et al., 1996), a decrease not achievedat all in our patients.

Coronary blood flow

We did not observe any differences in the qualitative assessment of the blood flowvelocity profile between orientation or type of valve prostheses. In several studiescoronary blood flow velocity profile was investigated and the effect of AVR in patientswith an aortic valve stenosis (Fujiwara et al., 1989; Kenny et al., 1994; Petropoulakiset al., 1995; Fujiwara, 2001; Takeuchi and Nakashima, 1997). As opposed to normalcoronary flow profile in resting conditions, patients with an aortic valve stenosisshow a decreased systolic component of the coronary flow signal whereas even aretrograde signal is observed during early systole. This distinct flow profile is amongothers ascribed to the increased intramyocardial compressive forces in these patientsdue to an elevated transvalvular pressure gradient (Spaan, 1991). Intramyocardialcompressive forces play an increasingly important role in the phasic blood flowprofile during hemodynamic stress or exercise. In one of the patients participatingin our study the circumstances allowed a blood flow velocity measurement in restingconditions and a subsequent measurement during maximal adrenergic stimulation.These measurements are shown in figure 4.4. It can be observed that in restingconditions an antegrade systolic component is present figure 4.4 panel A whichbecomes partly retrograde during administration of dobutamine figure 4.4 panel B.Compressive forces become more pronounced due to, on the one hand, increasedcontractility (Petropoulakis et al., 1995) and on the other hand transvalvular pressuregradient increases as a result of an increased cardiac output during dobutamineinfusion (Fiore et al., 2002).

42 Chapter 4

Transvalvular pressure gradient

Besides differences in coronary flow, Kleine et al. (Kleine et al., 2002b) also showeddifferences in hemodynamic performance for different orientations of mechanicalvalve prostheses. They found that transvalvular pressure gradients in the ’optimal’orientation, corresponding to the orthotopic orientation A in our study, was lowercompared to orientation B. However, we found no differences in transvalvularpressure gradients with echocardiography between the two orientations. It istherefore unlikely that intramyocardial compressive forces and also coronary bloodflow, as a consequence of differences in transvalvular pressure gradients, are differentfor the two orientations of the mechanical bi-leaflet valve prosthesis.

4.4.1 Limitations

In this in-vivo study we did not measure transvalvular pressure drop in combina-tion with the accompanying intra-myocardial compressive forces or microvascularresistance during catheterisation. To do so, simultaneous intra-ventricular pressuremeasurement would also have been mandatory, which for obvious reasons couldnot be performed. Furthermore, because accurate absolute flow measurement incoronary arteries is not possible in conscious humans neither invasively or non-invasively with MRI, we used coronary perfusion pressure as a surrogate of flow. Thisis justified because at maximum hyperemia resistance in the coronary circulation isminimal and constant and blood flow is directly proportional to perfusion pressure(Pijls, 2004; Pijls et al., 1996, 1993). Finally, the number of patients in our studywas small, but since the protocol was extensive we consider the number of patientsin this pathophysiologic study sufficient especially since the pressure differences weobserved were small and clinically irrelevant.

4.5 Conclusion

The influence of a bi-leaflet prosthesis on coronary perfusion pressure is negligibleand not dependent on valve orientation. Moreover, coronary perfusion pressures donot differ significantly between mechanical and biological prostheses.

Acknowledgements

This study was supported by an educational fund of Medtronic Inc.. The authors areindebted to the nursing staff of the catheterisation and echocardiography laboratoryof the Catharina Hospital Eindhoven for their dedicated assistance in performing allprocedures.

Chapter 5

Continuous infusionthermodilution for assessment

of coronary flow: Theoreticalbackground and in-vitro

validation

Direct volumetric assessment of coronary flow during cardiac catheterisation has not been availableso far. In the present study continuous infusion thermodilution, a method based on continuousinfusion of saline into a selective coronary artery, is evaluated. Theoretically, volumetric flowcan be calculated from the known infusion rate (Qi), the temperatures of the blood (Tb), thesaline (Ti), and the mixture downstream of the infusion site (T). We aimed to validate andoptimize the measurement method in an in-vitro model of the coronary circulation. Full mixing ofinfusion fluid and blood was found to be the main prerequisite for accurate determination of thecoronary flow. To achieve full mixing the influence of catheter design, infusion rate, and locationof temperature measurement were assessed. We found that continuous infusion thermodilutionslightly overestimated coronary flow determined by directly measured reference flow by 7±8%,over the entire physiologic flow range of 50-250 ml/min. These results were found using a speciallydesigned infusion catheter (infusion mainly through distally located side holes), a high enoughinfusion rate (25 ml/min), and measurement of the mixing temperature between 5 and 8 cmdistal from the tip of the infusion catheter. Absolute coronary flow rate can be measured reliablyby the continuous infusion method when full mixing is present, under the conditions mentionedabove.

The contents of this chapter are based on: M.van ’t Veer∗, M.C.F. Geven∗, M.C.M. Rutten,A. van der Horst, W.H. Aarnoudse, N.H.J. Pijls, and F.N. van de Vosse. Continuous infusionthermodilution for assessment of coronary flow: theoretical background and in-vitro validation.Medical Engineering and Physics, submitted]

∗ Both authors contributed equally to the study. ] Parts also printed in Geven, PhD thesis (2007), chapter 4

43

44 Chapter 5

5.1 Introduction

In the assessment of the coronary circulation intracoronary pressure and blood floware the parameters characterizing the functional significance of disease. Intracoronarypressure is widely used to quantify the severity of an epicardial stenosis (de Bruyneet al., 1996; Pijls, 2004). To assess the condition of the myocardial microvasculature,however, also quantification of absolute coronary and myocardial flow is needed(Aarnoudse et al., 2004b; Fearon et al., 2003).

Techniques for direct absolute blood flow measurement are not available forcommon clinical practice in the catheterisation laboratory. Therefore, indirectmeasures are used for the determination of coronary or myocardial blood flow,such as blood flow velocity or transit times (White, 1993; Meier and Zierler, 1954).During catheterisation blood flow velocity measurements may be carried out using anultrasound Doppler-crystal, mounted on a guide wire. Blood flow velocity is recordedand consequently absolute flow (flow rate) is estimated assuming a Poiseuille profileand integrating over the vessel’s cross section. Relatively large errors up to 20% aremade using this technique (Buchi and Jenni, 1998), whereas in a considerable partof the patients (up to 35%) no reliable measurement can be obtained (Barbato et al.,2004).

Another invasive method uses the injection of an indicator into the blood, andmonitoring the transit time of this indicator in the blood flow. The conventionalclinically applied indicator dilution method is thermodilution, where the indicator isa bolus of saline at room temperature, briskly injected into the coronary ostium. Themean transit time of the bolus is inversely correlated to the coronary flow. However,still no absolute flow rate can be measured unless the exact vascular volume is known(Barbato et al., 2004; de Bruyne et al., 2001c; Pijls et al., 2002a).

Theoretically, absolute blood flow can be measured from the mixing temperatureof a known infusion rate at known temperature, and the constant temperature of theblood. Continuous infusion thermodilution was proposed by Ganz et al. (Ganz et al.,1971) more than 30 years ago, for the measurement of blood flow in the coronarysinus. However, besides the fact that such coronary sinus measurements could notdifferentiate between blood flow from the different coronary arteries and differentmyocardial territories, the variability was too high to be useful for clinical applicationand the methodology was soon abandoned (Mathey et al., 1978; Weisse and Regan,1974). Recently, we applied continuous infusion thermodilution in animal and patientstudies to determine absolute coronary blood flow in a select coronary artery duringcardiac catheterisation (Aarnoudse et al., 2007) (chapter 6 and 7).

In these studies, we found strong correlations between real coronary flow andthe flow determined by the continuous infusion thermodilution method. In analyzingthe data we hypothesized that complete mixing of the infused fluid and the bloodcomprised the main prerequisite for applicability of the continuous infusion methodfor measurement of absolute coronary blood flow. The design and characteristics ofthe infusion catheter appeared to be important.

In the current study we provide a more detailed theoretical background, and takethe method to the bench to investigate its fundamental characteristics more closely

Coronary flow measurement by continuous infusion thermodilution: In-vitro study 45

under well-controlled conditions in a physiologically representative model of thecoronary circulation (Aarnoudse et al., 2004b; Geven et al., 2004). The model allowsfor measurement and control of all relevant parameters such as fluid temperature andcoronary flow rate.

The aim of the present study is to investigate the boundary conditions for optimalmixing and accurate application of the method with respect to the design of theinfusion catheter, different infusion rates, and the sites for measurement. Hereto,the flow rates derived by this method are compared with real flow, obtained with aperivascular ultrasonic flow probe.

5.2 Methodology

5.2.1 Theoretical background and measurement principle

The temperature distribution in the vessel during infusion can be described using theheat equation, expressing the conservation of energy within the system. The generalform is given by:

ρcp[∂T

∂t+ (~u · ∇)T ] = k∇2T (5.1)

where ρ is density, cp is specific heat, T is temperature, t is time, u is velocity,and k is thermal conductivity. The first term on the left hand side describes thelocal temperature variations in time. In the continuous infusion experiments the heattransfer due to mixing is assumed to be stationary, hence ∂T

∂t = 0. The diffusion ofheat is assumed to be small compared to convective heat transfer due to the fluidflow: (~u · ∇)T � k∇2T . Now the simplified heat equation becomes:

ρcp(~u · ∇)T = 0 (5.2)

In this general form, only one fluid is considered. In the measurement situationcold fluid (subscript i) is added to the blood (subscript b):

ρbcp,b(~ub · ∇)T + ρicp,i(~ui · ∇)T = 0 (5.3)

When optimal mixing is assumed, integration over the vessel surface (to achievevolumetric flow Q) and in the direction of the flow (to derive temperature differencesfrom temperature gradients) is allowed and gives:

ρbcp,bQb(Tb − T ) + ρicp,iQi(Ti − T ) = 0 (5.4)

By rearranging (5.4) the expression for Qb can be found:

Qb =ρicp,i(T − Ti)ρbcp,b(Tb − T )

Qi =ρicp,i

ρbcp,b

[Tb − Ti

Tb − T− 1

]Qi (5.5)

where ρb and ρi are the densities of the blood and the indicator and cp,b and cp,i

are the specific heats of the blood and the indicator respectively.

46 Chapter 5

In equation 5.5, Qb is the blood flow during infusion of the indicator, usuallysaline, and is assumed not to be affected by the infusion. However, if aortic pressureis not increased by the infusion and the myocardial resistance remains constant, thetotal flow through the myocardium will not increase. Hence, Qb is affected and partof the blood flow will be replaced by the infusion rate. The measured flow duringinfusion is decreased by Qi and the original blood flow before infusion can be foundusing:

Qb,orig =ρicp,i

ρbcp,b

[Tb − Ti

Tb − T− 1

]Qi + Qi (5.6)

In the remainder of this thesis Qth will be used instead of Qb,orig.

In the current in-vitro study water is used both as the flow medium (instead ofblood) and for the infused fluid (instead of saline) resulting in a value for the specificheat density-fraction of 1.002.

As mentioned, the principal condition for this technique requires for the infusedfluid and the flow medium to be fully mixed. A non-homogeneous mixture mightlead to a fluctuating mixing temperature measured distal from the site of infusionresulting in either an over- or underestimation of the coronary flow rate. The designof the infusion catheter influences the mixing process.

Second, the mixing temperature is assumed to be only determined by the twoinflow fluids, flow medium and infused fluid, and is not influenced by heat transferbetween the mixed fluid and the wall. Theoretically, the fluid will be heated graduallywith increasing distance from the infusion site.

Third, in the calculation of blood flow from fluid temperatures an assumption forthe possible change in myocardial resistance should be made: hence the influenceof the infusion to total myocardial flow should be determined. These boundaryconditions are studied in the in-vitro experiments.

5.2.2 In-vitro model and instrumental set-up

A full description of the physiologic representative experimental model we usedis described elsewhere (Geven et al., 2004; Aarnoudse et al., 2004b). In short,the model consisted of a piston pump, a left ventricular chamber and two valves,representing the left ventricle of the heart, a systemic and a coronary circulation.The systemic circulation contained a polyurethane tube (with the dimensions andmechanical properties of the aorta), and a system of compliances and resistances,creating physiologic aortic pressure and flow patterns. A polyurethane coronaryartery branched off the aorta directly distal to the aortic valve and bifurcates inepicardial branch and a sub-endocardial branch (figure 5.1, top panel). The sub-endocardial branch was led through the left ventricular chamber and collapsed duringsystole resulting in the typical physiologic coronary flow signal. A perivascularultrasound flow probe (4PSB, Transonic) was placed around the main branch ofthe coronary artery to measure true coronary flow (figure 5.1, bottom panel). The


arteriolar resistance was tuned to obtain hyperemic coronary flow of approximately250 ml/min for all measurements. A coronary stenosis was created by a clamp directlydistal to the flow probe, allowing for variation of coronary flow between 50 and250 ml/min. The model was submerged in water which was kept at a constanttemperature of 37.00±0.05◦C by an external thermal bath and circulator (F34-HL,Julabo).

The instrumental set-up for the continuous infusion experiments is depicted infigure 5.1, bottom panel. Aortic pressure was measured directly distal to the aorticvalve using a pressure transducer (P10EZ-1, Becton Dickinson) and bridge amplifier(Picas-CA2CF, Peekel Instruments). The guiding catheter was positioned near theostium of the coronary artery. A sensor tipped guide wire (PressureWire-5, RADIMedical Systems) was advanced through the guiding catheter into the coronary arteryto measure coronary pressure and temperature distal to the site where indicator wasinfused by the infusion catheter. The sensor of the guide wire is located 3 cm proximalto its floppy tip. Thereafter, the infusion catheter was advanced over the guide-wireand positioned in the coronary artery proximal to the site of the stenosis. The infusioncatheter was then connected by a Y-connector to the infusion pump (Angiomat 6000,Liebel-Flarsheim, Germany) and the guide wire was connected to the RADI Analyzer(RADI Medical Systems).

5.2.3 Measurement protocol

Infusion catheter and infusion rate

Two different over-the-wire infusion catheters were evaluated: the first one wasa general model frequently used in the catheterisation laboratory (model A), withthree side holes equally distributed over the distal 3 cm of the catheter, the second aspecifically designed catheter (model B), in which four side holes were laser-punchedover a length of 0.5 cm, from 0.5 to 1.0 cm from the tip. The distal end of catheterB was tapered to minimize infusion through the end-hole. Geometrical properties ofthe catheters are detailed in table 5.1. The most distal side hole was positioned 5mm from the tip of both catheters. Care was taken to locate the tip of the infusioncatheter such that all side holes were positioned inside the coronary artery, but closeto the coronary ostium. Mixing is facilitated by non-fully developed entrance flowand secondary flows due to the curvature and irregular shape of the connection.

Measurements were carried out at two infusion rates: a low rate of 15 ml/minand a higher rate of 25 ml/min.

Temperature registration during guide wire pullback

To investigate the mixing behavior of the infused fluid with the heated water in themodel, the influence of warming of the mixture via the vessel wall, and the optimallocation for measurement of the temperature of the infused fluid Ti, a pullbackprotocol was carried out. The local temperature during infusion was measured bystep-wise manual pullback of the guide wire, from a sensor position 10 cm distal tothe tip of the infusion catheter to 4 cm inside the infusion catheter at intervals of 1

48 Chapter 5

Figure 5.1: Physiologic representative model with magnification of the instrumentalset-up (main panel). Left ventricle is denoted by LV. Inset: The infusion catheterand the sensor tipped guide wire are positioned in the coronary artery through theguiding catheter. Indicator is infused (Qi) and temperatures are measured before (Tb)and during infusion (T). The guide is pulled back into the infusion catheter to obtainTi. True coronary flow is measured by the ultrasound flow probe and coronary flowis adjusted by the clamp.


Table 5.1: Geometrical properties of the infusion catheters.Model A Model B

Number of side holes 3 4Side holes distributed over 30 mm 5 mmOuter diameter 1 mm 1.17mmInner diameter 0.53 mm 0.97 mm

cm. Pullback temperature curves were made for both catheters at coronary flow ratesof 50 and 250 ml/min.

Flow measurement at fixed position

Coronary flow was derived by continuous infusion thermodilution (Qth) usingequation 5.6 and was correlated to the real flow measured by the ultrasound flowprobe (Qmeas). The coronary flow rate was varied from 50 to 250 ml/min in steps of50 ml/min. From the pullback curves the optimal sensor positions for measurementof mixing temperature T and infusion temperature Ti were derived. Aortic pressure,temperature, and coronary pressure and flow were monitored. The sensor waspositioned distal to the tip of the infusion catheter at the location determined by theguide wire pullback experiment. Infusion was started at a rate of Qi and the mixingtemperature T was registered, subsequently the guide wire was pulled back into theinfusion catheter for the measurement of Ti. Data were analyzed using Bland-Altmanplots (Bland and Altman, 1986).

5.3 Results

5.3.1 Temperature course of guide wire pullback

All temperature registrations during the pullback curves show variations in tem-perature between 3 and 10 cm from the catheter tip for both infusion cathetersand infusion rates (figure 5.2(a)) for both catheters. The instationary variations intemperature between 0 and 3 cm from the tip were larger than further downstream.To evaluate the course of the temperature difference as a function of the distance,all temperatures were normalized to the mean temperature between 3 and 10 cm(figure 5.2(b)). The shape of the curves was similar for both catheters and infusionrates and the distance could be divided into three characteristic zones, (a) <5 cm: thetemperature was unsteady (model A) or decreasing (model B) towards the tip, (b) 5 -8 cm: a steady or decreasing temperature was present with increasing distance distalfrom the tip, (c) > 8 cm: a gradual rise in temperature was seen with both catheters.

5.3.2 Flow measurement at fixed position

For the calculation of the coronary flow the myocardial resistance was assumed to beconstant before and during infusion. This was confirmed by the registration of the

50 Chapter 5

−6 −4 −2 0 2 4 6 8 10−10

−8

−6

−4

−2

0

distance to tip [cm]

∆T [° C

]

catheter A

−6 −4 −2 0 2 4 6 8 10−10

−8

−6

−4

−2

0


∆T [° C

]

catheter B

inside catheter mixing area measure distal

(a) Temperature registration during pullback of the sensor.

2 3 4 5 6 7 8 9 10 11

0.8

0.9

1

1.1

1.2


norm

aliz

ed ∆

T [−

] catheter A

mixing measure distal

2 3 4 5 6 7 8 9 10 11

0.8

0.9

1

1.1

1.2


norm

aliz

ed ∆

T [−

] catheter B

(b) Normalized temperature registration during pullback of the sensor outside the catheter.

Figure 5.2: Temperature registration during pullback of the sensor. Two infusioncatheters were used: model A (top) and model B (bottom) of both (a) and (b).Figure (a) shows the temperature at several locations measured with respect to thetemperature of the flow medium (Tb). In figure (b) temperature difference (Tb-T) isnormalized to the mean temperature difference over 3-10 cm, distal to the infusionsite for both catheters. Open symbols denote coronary flow of 250 ml/min, filledsymbols 50 ml/min, infusion rates: � = 25 ml/min, o = 15 ml/min. The distanceis measured relative to the tip of the infusion catheter. Catheter specifications aresummarized in table 5.1.


real coronary flow measured by the ultrasound flow probe: the relative differencein flow between before and during infusion was 0.8±1.1 % for all fixed positionmeasurements. Moreover, pressure did not increase during infusion. Hence, theoriginal coronary flow was calculated by equation 5.6.

From the temperature course measurements the positions to measure mixingtemperature T and infusion temperature Ti were found to be between 5 and 6 cmdistal to the tip of the infusion catheter and 1 cm within the catheter respectively.At a coronary flow of 50 ml/min, for both infusion rates and both catheters the flowrate was accurately determined (figure 5.3). At an infusion rate of 15 ml/min, forboth catheters increasing flow rate was increasingly underestimated (figure 5.3(a)and 5.3(c)). Hence, the mixing temperature was too low. At the higher infusionrate of 25 ml/min, the flow rate was reliably calculated by the continuous infusionmethod with catheter B (figure 5.3(d)), whereas the difference between calculatedand real flow deviates from zero for coronary flow rates > 200 ml/min for catheter A(figure 5.3(b)). For consistent flow estimation catheter B should be used at the highinfusion rate of 25 ml/min.

The accuracy of the method could be determined, provided that no relationbetween the absolute value and the difference between the real and calculated flowwas present. In the Bland Altman plots for catheter A at both infusion rates and forcatheter B at low infusion rate, such a relation is observed (figure 5.3(a)-(c)). Theabsolute flow rate is progressively underestimated with increasing flow rate. Thisindicates that the accuracy of the measurement is strongly related to the absolutevalue of the coronary flow. A general estimation for the accuracy can thereforenot be given. Catheter B used in combination with a high infusion rate resultedin a slight overestimation of 7±8 % of true coronary flow which corresponded toa mean absolute difference between calculated and direct measured flow over theentire range of 7±4 ml/min (figure 5.3(d)).

5.4 Discussion

An in-vitro model was used to investigate the most critical boundary condition forclinical application of continuous infusion flow measurement: the mixing of the flowmedium and infused fluid.

The mixing is influenced by infusion catheter design, position of the temperaturesensor, and infusion flow rate. The effects of variation in these parameters on thedetermination of the coronary flow were investigated. From the guide wire pullbacka position 5 to 6 cm distal to the tip of the infusion catheter appeared to be aproper distance to obtain the relevant mixing temperature T. However, a varyingtemperature with the distance to the tip of the infusion catheter was found, whichindicates incomplete mixing. The pattern of the variations was similar for bothinfusion catheters and both flow rates used. The first few centimeters distal to thetip may be defined as mixing chamber, from 5 to 8 cm the temperature was relativelystable, and after 8 cm the temperature difference between the mixture and inflowblood decreased. The latter effect may be due to heat loss through the wall. However,

52 Chapter 5

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qmeas

ml/min

Qth

ml/m

in

A − 15 ml/min

(a)

0 50 100 150 200 250 300−50

−40

−30

−20

−10

0

10

20

30

40

50

Average Qth

and Qmeas

[ml/min]

Per

c. d

iff. Q

th a

nd Q

mea

s [%]

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qmeas

ml/min

Qth

ml/m

in

A − 25 ml/min

(b)

0 50 100 150 200 250 300−50

−40

−30

−20

−10

0

10

20

30

40

50

Average Qth

and Qmeas

[ml/min]

Per

c. d

iff. Q

th a

nd Q

mea

s [%]

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qmeas

ml/min

Qth

ml/m

in

B − 15 ml/minB − 15 ml/min

(c)

0 50 100 150 200 250 300−50

−40

−30

−20

−10

0

10

20

30

40

50

Average Qth

and Qmeas

[ml/min]

Per

c. d

iff. Q

th a

nd Q

mea

s [%]

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qmeas

ml/min

Qth

ml/m

in

B − 25 ml/minB − 25 ml/minB − 25 ml/min

(d)

0 50 100 150 200 250 300−50

−40

−30

−20

−10

0

10

20

30

40

50

Average Qth

and Qmeas

[ml/min]

Per

c. d

iff. Q

th a

nd Q

mea

s [%]

Figure 5.3: Comparison of the calculated flow Qth to the directly measured flowQmeas for the two catheters A and B and for infusion rates of 15 ml/min and 25ml/min. Results for catheter A are shown in (a) and (b) and the results for catheter Bin (c) and (d). On the right hand side the relative differences between the techniquesare plotted against the average of the two techniques.


no clear reason for the characteristic 8 cm for this effect to occur is available. Analternative explanation for the varying temperature over the range from 3 to 10 cmmay be found in the presence of a swirl of the cold infused fluid in the vessel. In thatcase the reliability of the method is negatively influenced, because no guarantee forcomplete mixing can be given then. It is less likely that this effect occurs in a coronaryartery, because of the large variety in curvature of the vessels and the motion of thecoronary arteries due to the beating of the heart. The temperature of the infusedfluid Ti is determined inside the infusion catheter during the procedure. For catheterA small variations in Ti were observed, whereas for catheter B variations were larger.However, the variations in Ti did not significantly influence the calculations, becauseof the limited relative importance of a small absolute variation in Ti to the ratio ofT and Ti . Therefore, 1 cm within both catheters was found to be an adequate locationto obtain Ti.

The mixing became worse with increasing coronary flow rates reflected in lower(more negative) mixing temperatures of T. Consequently, these lower values translateinto lower calculated coronary flow values. This observation may be explained by thedecreasing ratio of infusion flow velocity and coronary flow velocity. The presenceof side holes, and their size and position, determine the inflow velocity and directionof the infused fluid. The side holes positioned closely together in combination witha tapered end hole are factors which we expected to enhance mixing due to the”spraying” effect during infusion. The measurement with catheter B was the mostaccurate over the entire range for the high infusion rate of 25 ml/min (figure 5.3),confirming our hypothesis. The influence of the infusion velocity on the flow mighthave been damped out between the side holes of catheter A, which were relatively farapart. Moreover, infused fluid infused through the open end hole, in the direction ofthe main coronary flow, would not have facilitated mixing, also explaining the inferiorresults of catheter A.

However, when the position of the tip of the infusion catheter B was only slightlyadvanced into the coronary artery by 3 cm, mixing was immediately worse and theflow was underestimated (data not shown). Thus, the occurrence of secondary flowsor flow disturbances, resulting from the inflow which is not fully developed in theirregularly shaped modeled coronary ostium, is of great importance for the mixing totake place.

Proper mixing is the main determinant for the technique to be useful. The smoothnature of the coronary artery in the in-vitro model limits formation of complex flowpatterns which consequently complicates mixing. The complex geometry of the bloodvessel in combination with the beating of the heart will enhance proper mixing in thein-vivo situation and hence more accurate flow calculations may be obtained. On theother hand, water is used as a medium in our in-vitro set-up, while the flow was notscaled, so the relatively high Reynolds number could indicate an increased probabilityfor flow instabilities to occur, enhancing the mixing process.

The difference in viscosity between water and blood may additionally affect themixing process. Also the heat transfer to the wall may be different between the waterin the model tube and the blood in the vessel. However, when the sensor position ischosen close to the end of the mixing zone, no significant influence is expected.

54 Chapter 5

5.5 Conclusion

Absolute coronary flow rate can be directly measured reliably over the entirephysiologic range of 50 to 250 ml/min by the continuous infusion method, underthe condition that a suitable infusion catheter is used at a high infusion rate of 25ml/min and positioned in an area with a complex flow pattern. The measurementlocation should be chosen appropriately, slightly distal to the mixing zone.

5.5.1 Clinical implications

Because clinically applicable catheters and infusion equipment were used, theoptimization of the methodology and catheter design performed in this study canbe directly used in the catheterisation laboratory. Optimal mixing between the salineand the blood was indicated to be the main prerequisite for successful measurementof coronary flow. The nature of coronary flow in-vivo is expected to be even morefavorable towards optimal mixing than the in-vitro flow, due to the complex geometryof the arteries and the beating of the heart. Therefore, this study contributes to thefeasibility of continuous infusion flow measurement in selective coronary arteries,providing the first methodology to perform such absolute flow measurement.

Acknowledgements

This study was supported by the Dutch Technology Foundation (STW) projectEPG.5454, and by RADI Medical Systems, Uppsala, Sweden. The authors would liketo thank Boston Scientific, Natick (MA), USA, and Occam International, Eindhoven,The Netherlands, for the design and production of the infusion catheters.

Chapter 6

Continuous infusionthermodilution for assessmentof coronary flow: Animal study

Based upon the principle of thermodilution with continuous low rate infusion of saline at roomtemperature, an instrumental set-up was designed for direct coronary blood flow measurementduring cardiac catheterisation. A 2.8 F infusion catheter was developed enabling selective infusionof saline into a coronary artery. Using this specially designed infusion catheter and a standard0.014” sensor-tipped pressure/temperature guidewire, absolute coronary blood flow in a coronaryartery could be calculated from the infusion rate of saline, temperature of the saline at the tipof the infusion catheter, and distal blood temperature during infusion. In the former chapterthe method was tested over a wide range of flow rates in an in-vitro set-up. In this chapter theinfluence of sensor position and infusion rate on agreement and reproducibility, and the safety ofthe technique was assessed in 5 chronically instrumented dogs. Using a suitable infusion catheterand a 0.014” sensor-tipped guidewire for measurement of coronary pressure and temperature,volumetric blood flow can directly and safely be measured in selective coronary arteries duringcardiac catheterisation.

Published in part in: W.H. Aarnoudse∗, M. van ’t Veer∗, N.H.J. Pijls, J. ter Woorst, S.Vercauteren, P.A.L. Tonino, M.C.F. Geven, M.C.M. Rutten, E. van Hagen, B. DeBruyne, and F.N.van de Vosse. Direct volumetric blood flow measurement in coronary arteries by thermodilutionJournal of the American College of Cardiology, 50:2294-2304, 2007.

∗ Both authors contributed equally to the study.

55

56 Chapter 6

6.1 Introduction

For the understanding of coronary physiology, several specific pathologic conditions,and evaluation of treatment, direct volumetric measurement of coronary blood flowin selective coronary arteries would be useful. Despite numerous efforts, directvolumetric measurement of blood flow in selective coronary arteries during cardiaccatheterisation has not been possible so far. Therefore, either non-invasive methods orsurrogate invasive measures have been used for that purpose, with all their inherentlimitations and inaccuracy (Rutishauser et al., 1967; Ganz et al., 1971; Wilson et al.,1985).

Volumetric blood flow measurement by thermodilution techniques and continuousinfusion of saline was proposed by Ganz et al. as early as 1971 but could only beapplied in the coronary sinus (Ganz et al., 1971). Theoretically, absolute blood flowin the coronary sinus was measured from a known infusion rate of saline at a knowntemperature and from blood temperature measured downstream (Ganz et al., 1971;Ziehler, 1962). Besides the fact that such a measurement could not differentiatebetween blood flow from the different coronary arteries and different myocardialterritories, its variability was too high to be useful for clinical application and themethodology was soon abandoned (Weisse and Regan, 1974; Mathey et al., 1978).Due to technical limitations of the equipment at that time the technique was appliedto the coronary sinus only. Measurements in selective coronary arteries were neverperformed, neither in animals nor in man.

In this chapter, we describe the experimental validation of a methodology basedupon the theory described by Ganz et al. (1971). Measurement of maximum coronaryblood flow in individual coronary arteries is performed during catheterisation.

6.2 Theoretical background and aim

As derived in chapter 5 coronary blood flow can be determined reliably from theblood temperature (Tb), temperature of the infused saline (Ti), the temperature of themixture of blood with saline (T), and the infusion rate (Qi) by equation 5.6 providedthat blood and saline are properly mixed.

This equation holds under the assumptions that heat transfer due to mixing isstationary, blood and saline are completely mixed at the site were T is measured,the diffusion of heat is assumed to be small compared to convective heat transferdue to the fluid flow, resistance of the myocardium is minimal and constant duringmeasurements, and coronary pressure is not increased during infusion of saline.Consequently, taking the differences in specific heat and density of the blood andthe saline into account and express T and Ti as the deviation with respect to theblood temperature (Tb) the original blood flow can be calculated using:

Qth = 1.08[Ti

T− 1

]Qi + Qi (6.1)

In chapter 5 the thermodilution method was taken to the bench to investigate

Coronary flow measurement by continuous infusion thermodilution: Animal study 57

the relevant parameters that influence the proper application of the method in aphysiologic well-controlled environment. In the current chapter the influence of thesensor position and infusion rate on agreement and reproducibility as well as thesafety of the technique itself will be investigated in five chronically instrumenteddogs.


6.3.1 Animal instrumentation

Five mongrel dogs (weight 28-42 kg) were used for this study. Animal handling wasperformed according to the European directive on laboratory animals (86/609/EEC)and the study protocol was approved by the animal ethics committee of the Universityof Maastricht. After premedication with acetadon 0,1 ml/kg intra muscularly andbuprenorfine 0,01 mg/kg intra venously, the dogs were anesthetized with thiopental15 mg/kg intra venously and ventilated by room air with 1% isoflurane and 0,5 lN2O per minute. A left thoracotomy was performed and the proximal left circumflexartery was dissected free. A perivascular volumetric flow meter of appropriate size(PS series, Transonic Inc, Ithaca, NY) was placed around the artery and a circularballoon occluder (IVM, Healdsburg, CA) was placed just distal to the flow probe forcreating stenosis of variable degree (figure 6.1). After the equipment was checked forproper functioning, the pericardium and chest were closed and the instrumentationleads were stored in a subcutaneous pocket in the neck of the animal until the timeof the study. Long acting ampicilin was administered one hour prior to the operationand repeated every 48 hours.

6.3.2 Cardiac catheterisation

Eight days after instrumentation, each dog was anesthetized again by nicomorfine10 mg/h intra venously and 1% isoflurane. The subcutaneous pocket was opened,the lead of the flow probe was connected to the appropriate recording equipment(Transonic 400 series Modular Flowmeter), and the tube of the occluder connectedto a 5 cc syringe for creating different coronary stenoses. Subsequently, cardiaccatheterisation was performed. A 6F arterial sheath and a 5F venous sheath wereintroduced into the left femoral artery and vein and a 6F left Judkins guiding catheterwas advanced into the ostium of the left coronary artery. Next, a commerciallyavailable 0.014” pressure/temperature sensor-tipped guidewire (PressureWire 5,Radi Medical Systems, Uppsala, Sweden) was introduced by a Y-connector andadvanced through the guiding catheter into the left circumflex artery and positionedwith its sensor approximately 6 cm distal to the flow probe and occluder. This sensor-tipped guidewire measures pressure with an accuracy of 1 mmHg at a frequency of600 Hz and temperature with an accuracy of 0.02◦C at a frequency of 600 Hz, andcan be connected to an appropriate interface (Radi-Analyzer, Uppsala, Sweden) forsimultaneous recording of aortic and coronary pressure and coronary temperature

58 Chapter 6

Figure 6.1: Instrumentation of the dogs and set-up during catheterisation. The leftcircumflex artery (LCX) is instrumented by a perivascular flow probe and occluder. A2.8 F infusion catheter is advanced over a 0.014” pressure/temperature sensor tippedguidewire through a Y-connector (Y1) and positioned with its tip just proximal to theflow probe and occluder. The infusion catheter is connected to an infusion pump by asecond Y connector (Y2), enabling continuous infusion of saline at room temperature(8 - 25 ml/min). The sensor-tipped guidewire is connected to the interface (RADIAnalyzer) as routinely done in coronary pressure measurement and distal coronarypressure (Pd) and temperature (T) are displayed on the interface. Also the aorticpressure (Pa), measured at the tip of the guiding catheter, is recorded by a regularpressure transducer and displayed on the interface (see figure 6.2).


(figure 6.2) (Aarnoudse et al., 2004a). Thereafter, a specially designed infusioncatheter was advanced over the sensor-tipped guidewire until its tip was just proximalto the flow probe. The details of this infusion catheter are described below. Theinfusion catheter was connected to a dedicated infusion pump (Angiomat 6000,Liebel-Flarsheim, Germany) by way of a second Y-connector. By doing so, saline couldbe infused through the infusion catheter while it was advanced over the guidewire(figure 6.1).

Infusion catheter

To ensure adequate mixing of blood and saline, a special infusion catheter wasdesigned based on findings from in-vitro experiments (see chapter 5). This catheterhad an outer diameter of 2.8 F, a tapered tip, an which closed gently around a 0.014”guidewire, and 4 small sideholes in the last 5 mm proximal to its tip (Occam Inc,Eindhoven, The Netherlands). The endhole of this infusion catheter precisely fittedthe guidewire in order to prevent saline from being infused through the endhole alongthe wire. When this infusion catheter was advanced across a guidewire, at least 40ml/min of saline at room temperature could be infused and left the distal end of thecatheter exclusively through the sideholes.

6.3.3 Experimental protocol

In every dog, measurements were performed in 4 series, corresponding to fourdifferent levels of a coronary stenosis: mild stenosis (i.e. balloon occluder notinflated at all), moderate stenosis, severe stenosis, and very severe stenosis. Totest reproducibility, all measurements were performed twice with an interval ofthree minutes. Furthermore, to test the independency of the methodology of theposition of the sensor, all measurements were performed at a distance of 6 and 3cm distal from the tip of the infusion catheter. To validate the independency of themethodology of the infusion rate, all measurements for all the degrees of stenosisand all sensor positions were also performed using a high and a low infusion rate (8- 15 and 15 - 25 ml/min, respectively, depending on the size of the coronary artery).In this way, a total of 4×2×2×2=32 measurements were performed in every dog.During the complete sequence of measurements, maximum hyperemia was inducedby continuous administration of adenosine 140 µg/(kg · min) in the femoral venoussheath, combined with administration of norepinefrine in a low dose, if necessary, tomaintain mean blood pressure at a level of 60-100 mmHg. The presence of steady-state maximum hyperemia was confirmed before and at the end of each series ofeight measurements by verifying that after a 20 seconds occlusion period, no furtherreactive hyperemia was present (Pijls et al., 1990).

6.3.4 Measurement procedure

During steady-state hyperemia, the blood temperature in the distal coronary artery(Tb), measured by the pressure/temperature sensor, was set to zero and the other

60 Chapter 6

temperatures were measured as deviation from this value. Thereafter, infusion ofsaline at room temperature was started at a constant rate (Qi; 8 - 25 ml/min) usingthe infusion pump. During steady-state continuous infusion of saline, the decreaseof temperature of the blood (T) after adequate mixing with the infused saline wasmeasured and recorded for 30 to 60 seconds (figure 6.2).

Next, the pressure/temperature sensor was pulled back into the infusion catheterand the temperature Ti of the infused saline at the location of the sideholes wasmeasured as deviation from Tb. Volumetric blood flow in the respective coronaryartery (Qth) was calculated then as described by equation 6.1. An example of such ameasurement and calculation is presented in figure 6.2.

Thereafter, the wire was advanced again to the original position and after waitingfor three minutes the procedure was repeated for assessing reproducibility. During theprocedure, all registrations were continuously displayed on the RADI-Analyzer, usingspecific software and a display, indicating T and Ti as the deviation from the bloodtemperature as measured immediately before saline infusion and making calculationsrather easy as mentioned above. During the complete procedure, also aortic pressure(Pa, recorded by the guiding catheter) and distal coronary pressure (Pd, recorded bythe sensor) were displayed continuously as well as fractional flow reserve (equation2.2) which is defined as Pd/Pa at maximum hyperemia (Pijls et al., 1993, 1995b).This sequence of in-duplo measurements was performed at the 4 different stages ofstenosis as described above.

Variation of infusion rate

Theoretically, the calculations should be independent of the infusion rate of saline(Ganz et al., 1971; Ziehler, 1962). Therefore, in all dogs and at every degree ofstenosis and at every sensor position, all measurements were performed with twodifferent infusion rates (low rate and high rate). The low rate varied between 8 and15 ml/min and the high rate between 15 - 25 ml/min, depending on the size of thecoronary artery. No infusion rates below 8 ml/min were used to avoid unfavorablesignal to noise ratios and no infusion rates above 25 ml/min were used to avoidexcessive cooling of the myocardium with possible adverse effects on the conductionsystem.

Variation of sensor position

Theoretically, the calculations should also be independent of the position of thetemperature sensor in relation to the infusion site (Ganz et al., 1971; Ziehler, 1962).Therefore, in all dogs at every degree of stenosis and at all infusion rates, themeasurements were performed with the sensor position at a distance of 3 cm and6 cm distal to the tip of the infusion catheter (proximal and distal sensor position,respectively). Positions less than 3 cm from the tip of the infusion catheter wereavoided to prevent measuring in an area with incomplete mixing, whereas a positiontoo far distally was avoided because of possible loss of indicator by inappropriateheating of saline by the vessel wall or surrounding myocardium.


Figure 6.2: Calculations of coronary blood flow in the LCX artery with a moderatestenosis. On the left, blood temperature at steady state hyperemia (Tb) is set to zero.Next, infusion of saline is started at a rate of 20 ml/min. During steady state salineinfusion, temperature in the coronary artery (T) decreases by 0.85◦C with respectto the blood temperature. Next, the sensor is pulled back to the tip of the infusioncatheter (asterisk) to measure Ti , which is 5.7 ◦C below initial blood temperature.Using equation 6.1 coronary blood flow can be calculated to be Qth= 143 ml/min.

62 Chapter 6


For all dogs, the relation between the calculated and measured blood flow wasevaluated by linear regression (Matlab, The Mathworks Inc., Natick, MA, USA).Agreement between the calculated and measured blood flow was assessed by Bland-Altman plots of the relative difference between Qth and Q (Bland and Altman, 1986).Also the reproducibility (first vs. second measurement), the influence of differentinfusion rates of saline (low vs. high rate), and the influence of sensor position (distalversus proximal) were evaluated by linear regression analysis and Bland-Altman plotsof relative differences. All hemodynamic data are given as mean±standard deviation.

6.4 Results

6.4.1 Hemodynamic characteristics and procedural results

The instrumentation was uneventful and uncomplicated in all dogs and the cardiaccatheterisation and measurements after eight days could also be performed withoutdifficulties in all 5 dogs. In the first dog, at the location of the coronary occluder, anangiographic stenosis of approximately 50% was visible and fractional flow reservewas 0.55, indicating that already a hemodynamically rather severe stenosis waspresent at the location of the instrumentation. Therefore, in this first dog only twoout of the 4 planned series of measurements could be performed (severe and verysevere stenosis). In all other dogs, the 4 complete series could be performed asplanned (mild, moderate, severe, and very severe stenosis). During all series in alldogs, steady state maximum hyperemia could be achieved by infusion of adenosine(140 µg/(kg · min) ) combined with a low dose of norepinephrine (0.10 - 0.30µg/(kg ·min) ). Mean arterial pressure, heart rate, and absolute blood flow measuredby the flow probe for the different dogs and the different series, are presented intable 6.1.

6.4.2 Flow measurements

Agreement

In all dogs, by different degrees of inflation of the perivascular balloon occluder,maximum blood flow could be varied over a wide range (table 6.1). The relationbetween blood flow as measured by the perivascular flow probe and blood flow ascalculated from the thermodilution experiments, is presented for all 5 individual dogsin figure 6.3. In all dogs except the first, an excellent correlation was found. Theagreement between calculated flow and measured flow for all dogs together as wellas the corresponding Bland-Altman diagram, is presented in figure figure 6.4(a).


0 50 100 1500

50

100

150Dog 1

Qmeas

[ml/min]

Qth

[ml/m

in]

0 50 100 150 200 250 3000

50

100

150

200

250

300Dog 2

Qmeas

[ml/min]

Qth

[ml/m

in]

0 50 100 150 200 250 3000

50

100

150

200

250

300Dog 3

Qmeas

[ml/min]

Qth

[ml/m

in]

0 50 100 150 200 250 3000

50

100

150

200

250

300Dog 4

Qmeas

[ml/min]

Qth

[ml/m

in]

0 50 100 150 200 250 3000

50

100

150

200

250

300Dog 5

Qmeas

[ml/min]

Qth

[ml/m

in]

Figure 6.3: Relation between blood flow as calculated by thermodilution (Qth) andtrue blood flow (Qmeas) in the individual dogs. Circles indicate a proximal sensorposition, squares a more distal sensor position (3 and 6 cm from the tip of the infusioncatheter, respectively). Closed symbols indicate high infusion rate of saline (15-25ml/min), open symbols indicate a low infusion rate (8-15 ml/min). The line indicatesthe line of identity. In dog1, only part of the measurements could be performed dueto a rather severe stenosis at the site of the instrumentation.

64 Chapter 6

Table6.1:

Hem

odynamic

dataofthe

fivedogs

atdifferent

seriesSten

osisSeverity:

Mild

Moderate

SevereVery

SevereBW

Pa

HR

Qm

ax

Pa

HR

Qm

ax

Pa

HR

Qm

ax

Pa

HR

Qm

ax

Dog

126

--

--

--

79136

10190

12454

Dog

235

79144

24282

128169

88130

12184

13765

Dog

334

77161

19667

128134

71142

11582

16257

Dog

441

83129

22491

127178

85127

12290

13172

Dog

530

72125

20158

155142

64151

9765

15044

BW=

Bodyw

eight(kg);H

R=

heartrate

(beats/min);P

a =m

eanblood

pressure(m

mH

g);Q

ma

x=

maxim

umcoronary

bloodflow

(ml/m

in.)


0 50 100 150 200 250 3000

50

100

150

200

250

300All measurements

Qmeas

[ml/min]

Qth

[ml/m

in]

(a)

High Inf, proxLow Inf, proxHigh Inf, distLow Inf, dist

y =0.73x+42R2 =0.72

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth

and Qmeas

[ml/min]

Rel

. diff

. Qth

and

Qm

eas [%

]

0 50 100 150 200 250 3000

50

100

150

200

250

300Reproducibility

Qth,1

[ml/min]

Qth

,2 [m

l/min

]

(b)

High Inf, proxLow Inf, proxHigh Inf, distLow Inf, dist

y =0.96x+3R2 =0.89

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,1

and Qth,2

[ml/min]

Rel

. diff

. Qth

,1 a

nd Q

th,2

[%]

0 50 100 150 200 250 3000

50

100

150

200

250

300High vs Low infusionrate

Qth,low

[ml/min]

Qth

,hig

h [ml/m

in]

(c)

y =1x+19R2 =0.66

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,low

and Qth,high

[ml/min]

Rel

. diff

. Qth

,low a

nd Q

th,h

igh [%

]

0 50 100 150 200 250 3000

50

100

150

200

250

300Proximal vs Distal position

Qth,prox

[ml/min]

Qth

,dis

t [ml/m

in]

(d)

y =1.1x − 4.7R2 =0.92

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,prox

and Qth,dist

[ml/min]

Rel

. diff

. Qth

,pro

x and

Qth

,dis

t [%]

Figure 6.4: Results for the animal experiments. Calculated versus measuredblood flow for all experiments in all dogs, represented by linear regression andcorresponding Bland Altman diagram (a). Second versus first measurement for allexperiments in all dogs (b). High versus low infusion rate (c). Distal versus proximallocation of sensor (d).

66 Chapter 6

Reproducibility

All measurements in all dogs, at all sensor locations and at all infusion rates, wereperformed twice with an interval of three minutes. The agreement between thefirst and second measurement is presented in figure 6.4(b). As can be observed,reproducibility was excellent. For all dogs together, Qth,2 equalled 0,96×Qth,1 +3 ml/min; R2 = 0.89. Also when studying reproducibility independently for thedifferent degrees of stenoses, the different sensor positions, or the different infusionrates, comparably high reproducibility was obtained.

Infusion rate

In all dogs and for all degrees of stenoses and all sensor positions, the in-duplomeasurements were performed at a low and high infusion rate. Infusion rate waschosen in such a way that distal coronary blood temperature at steady-state infusionwas in the range between 0.5 to 2.0◦C below blood temperature. Our experimentsrapidly taught us that in a large, almost normal coronary artery, the infusion rateshould be between 15 - 25 ml/min and in a stenotic artery 8 - 15 ml/min. In allour experiments we started with the lower infusion rate and increased this rate by afactor 2, called low and high infusion rate. In all dogs, a good agreement was foundbetween the calculated blood flow at both different infusion rates (Qth,high = 1.0 ×Qth,low + 19 ml/min; R2 =0.66). These results are presented in figure 6.4(c).

Position

In all dogs, at all degrees of stenosis and all different infusion rates, the in-duplomeasurements were performed at locations of 3 and 6 centimeter distal from thecoronary occluder. In all dogs and at all degrees of stenoses and all infusion ratesan excellent agreement was found between calculated blood flow at the 2 differentsensor positions (Qth,dist = 1.1 × Qth,prox - 5 ml/min; R2 = 0.92). These results aredisplayed in figure 6.4(d).

6.5 Discussion

We showed that the thermodilution technique for measuring volumetric blood flowin selective coronary arteries is feasible, reproducible, and safe in an in-vivo setting.In contrast to traditional coronary sinus thermodilution technique, this novel methodmakes it possible to selectively determine blood flow in individual coronary arteries.Its variability is smaller, instrumentation is less complicated and measurements can beperformed easily and quickly with standard equipment during coronary interventionsusing the diagnostic pressure/temperature guidewire, the corresponding interface(RADI Analyzer) and a suitable infusion catheter.

The superiority of our technique in the coronary arteries above traditionalcoronary sinus thermodilution (Ganz et al., 1971; Weisse and Regan, 1974; Matheyet al., 1978) is most likely explained by the better mixing of the blood and the saline.


Despite the retrograde infusion of saline used in the application in the coronary sinus,which would have had a positive effect on the mixing process compared to antegradeinfusion in our application, it is unlikely that blood and saline were completely mixedat the location where T was obtained. In our experiments we observed varyingtemperatures in the first few centimeters downstream of the infusion catheter anddefined this area as the mixing chamber (see chapter 5). In the traditional coronarysinus technique the temperature sensor was located downstream of the tip of theinfusion catheter which could explain the variability in the measurements. There areat least two reasons why our method is more accurate and has smaller variability thanthe measurements in the coronary sinus. First, we used a specially designed infusioncatheter with a tapered end hole and four side holes closely positioned together tofacilitate mixing, as indicated in chapter 5. Second, compared to the traditionalmeasurement in the coronary sinus, the more pronounced pulsatile motion of thecoronary artery, and consequently more complex flow profiles, also enhanced themixing.

Our results showed that the exact position of the sensor in the coronary artery,within certain limits, does not influence the calculated flow significantly. Comparableto our findings in the in-vitro experiments, between 3 and 6 cm distal to the tip of theinfusion catheter heating of the indicator through the arterial wall is negligible.

Also the amount of infused indicator can be varied within certain limits withoutaffecting the accuracy of the measurements, which is in accordance to theory (Ganzet al., 1971). From our in-vitro experiments better results were obtained formeasurements using high flow rates, whereas in the current animal study the infusionrate does not affect flow calculations. It seemed, however, that flow was more oftenoverestimated with high infusion rates compared to the measurements performedwith the low infusion rates.

6.5.1 Limitations

Although absolute coronary blood flow has been considered as the ultimate goalof coronary physiologists for decades, volumetric flow assessment in itself hasseveral limitations for practical purposes. First, absolute coronary blood flowcannot be interpreted without knowledge of the size of the perfused myocardialterritory (Gould, 1988; Bol et al., 1993; de Bruyne et al., 2001b). Second, it isdependent on hemodynamic variations in blood pressure, heart rate, and contractility(de Bruyne et al., 1996; Rossen and Winniford, 1993). Therefore, interpretationof a single value is rather difficult. However, because fractional flow reserve canbe measured simultaneously and relates a particular value of flow to its theoreticalnormal maximum value achievable in the respective coronary artery under similarhemodynamic conditions, interpretation of individual values of flow is made easier(Pijls et al., 1993). Third, it is not coronary but myocardial blood flow which ismost important for a patient, preferably calculated during maximum hyperemia. Innormal coronary arteries, it can be hypothesized that myocardial blood flow equalscoronary blood flow but in the presence of a stenosis, the collateral component playsan increasing role with increasing stenosis severity (Pijls et al., 1995a; Aarnoudse

68 Chapter 6

et al., 2004a).Theoretically, if the technique is used in patients with a very severe stenosis with

decreased resting coronary blood flow, blood flow can be so slow that indicator (cold)will be lost through the arterial wall, resulting in overestimation of true flow andmaking the methodology unreliable. In our series, however, this problem was notobserved despite the presence of severe lesions with an FFR as low as 0.21. Therefore,for clinical purposes, this limitation does not seem to be very relevant.

Finally, to ensure that blood flow was kept constant during the measurementitself, we only measured during maximum hyperemia. As explained in chapter 5,myocardial resistance reserve should be exhausted in order to use equation 5.6 andtherefore equation 6.1. If this situation is not achieved, autoregulation might causetotal flow to increase during infusion of saline. The presence of steady-state maximumhyperemia was confirmed before and at the end of each series of measurements byverifying that after a 20 seconds occlusion period, no further reactive hyperemiawas present (Pijls et al., 1990). Moreover, we checked if pressure during infusionincreased, which would cause an increase in flow as well. Since this was was not thecase, it was appropriate to use equation 6.1.

With this technique it is not possible to measure resting flow. This is not atrue limitation because maximal achievable blood flow is the most important clinicalparameter to characterize the coronary circulation and the true hemodynamic severityof an epicardial coronary stenosis (Pijls et al., 1993; Kern et al., 1997).


Being able to determine absolute blood flow in selective coronary arteries quantita-tively in a fairly simple and straightforward way during cardiac catheterisation, hasnot been possible so far and has several advantages: Not only coronary flow in itselfis measured, but because the guidewire also measures distal coronary pressure andfractional flow reserve, values of flow can be directly related to its normal maximumvalue and therefore better be interpreted.

Because also distal coronary pressure is measured simultaneously by the samesensor, myocardial resistance of a specific myocardial territory can be calculated.Hence, myocardial resistance equals myocardial perfusion pressure (measured by thesensor) divided by myocardial blood flow (Aarnoudse et al., 2004a).

So, we believe that this novel methodology for direct volumetric coronaryflow measurement can be a valuable tool in the clinical research of the coronarycirculation.

6.6 Conclusion

Continuous infusion thermodilution can be used safely during catheterisation toobtain maximal coronary blood flow in dogs. The methodology is feasible, repro-ducible and fairly accurate. Together with distal coronary pressure, measured by thesame guide wire, also absolute resistance of the coronary artery and the coronary


microcirculation can be calculated. Therefore, this methodology constitutes the firstdirect volumetric blood flow measurement in selective coronary arteries and willfacilitate pathophysiologic studies of the coronary circulation.

Acknowledgement

The authors wish to express their gratitude to the technicians and staff of the CentralAnimal Laboratory of the University of Maastricht for their hospitality and marveloussupport in performing the animal experiments.

Ethical considerations

Realistic coronary geometries in combination with the beating of the heart and theassessment of safety and applicability of the technique in an in-vivo situation limitedthe use of the in-vitro model. In chapter 5 the complex geometry of the blood vesselin combination with the beating of the heart in an in-vivo situation was assumed toenhance proper mixing and hence, more accurate flow calculations. So, the next stepin the development was to apply the technique in an in-vivo model. Because of theneed for direct validation of the blood flow, a flow probe had to be placed around acoronary artery. The latter operation could be performed in an animal model only.

In this study five dogs were used to validate the continuous infusion thermod-ilution technique in an in-vivo environment. Dogs were chosen because of theircomparable heart size and the extensive experience with the dog model. A dogweighing 20 to 40 kg has a comparable size of the heart as a 70 kg weighing man. Asa consequence coronary blood flow is in the same range.

A number of five animals was chosen to be able to assess if the technique wasdependent on the individual dog resulting from the biological diversity. Moreover,after five animals, in each of which 32 measurements were performed, enoughexperience would be present to continue the experiments in humans. No adverseeffects as a consequence of the continuous infusion thermodilution technique wereobserved.

70 Chapter 6

Chapter 7

Continuous infusionthermodilution for assessment

of coronary flow: Human study

Based upon the principle of thermodilution with continuous low rate infusion of saline at roomtemperature, an instrumental set-up was designed for direct coronary blood flow measurementduring cardiac catheterisation. A commercially available infusion catheter was used to selectivelyinfuse saline into a coronary artery. Using this infusion catheter and a standard 0.014” sensor-tipped pressure/temperature guidewire, absolute coronary blood flow in a coronary artery couldbe calculated from the infusion rate of saline, temperature of the saline at the tip of the infusioncatheter, and distal blood temperature during infusion. In the former chapter the method wastested over a wide range of flow rates in an in-vivo setting. In this chapter the influence ofsensor position and infusion rate on agreement and reproducibility, and the safety of the techniquewas assessed in thirty-five patients referred for elective percutaneous coronary intervention orintracoronary physiologic measurements. Using a suitable infusion catheter and a 0.014” sensor-tipped guidewire for measurement of coronary pressure and temperature, volumetric blood flowcan directly and safely be measured in selective coronary arteries during cardiac catheterisation inconscious humans.

Published in part in: W.H. Aarnoudse∗, M. van ’t Veer∗, N.H.J. Pijls, J. ter Woorst, S.Vercauteren, P.A.L. Tonino, M.C.F. Geven, M.C.M. Rutten, E. van Hagen, B. DeBruyne, and F.N.van de Vosse. Direct volumetric blood flow measurement in coronary arteries by thermodilutionJournal of the American College of Cardiology, 50:2294-2304, 2007.

∗ Both authors contributed equally to the study.

71

72 Chapter 7

7.1 Introduction

Direct volumetric blood flow measurement in selective coronary arteries in consciousman has not been possible so far. For assessing the physiologic significance of anepicardial stenosis, this is not a major problem, because other and easier indexeshave been developed for hemodynamic quantification of epicardial stenosis severity,such as fractional flow reserve (di Mario et al., 1995; Pijls et al., 1996; de Bruyneet al., 2001c). For the diagnosis and understanding of myocardial or microvasculardisease, however, absolute blood flow measurement would be a great step forwardbecause in such case, also absolute resistance can be quantified. Besides that, fromthe scientific point of view, direct measurement of volumetric coronary blood flowduring catheterisation has been a major goal for decades (Ganz et al., 1971).

In chapter 6 it was shown that coronary blood flow could be measured safelyduring catheterisation in dogs. The methodology showed to be safe, feasible,reproducible and fairly accurate in the in-vivo application. In this chapter thethermodilution technique is applied in conscious humans during catheterisation.When coronary blood flow is measured together with distal coronary pressure by thesame sensor-tipped guidewire, also absolute myocardial blood flow, collateral flow,and myocardial resistance can be calculated.

7.2 Theoretical background and aim

As derived in chapter 5 and confirmed in chapter 6 coronary blood flow can bedetermined reliably and reproducibly from the blood temperature (Tb), temperatureof the infused saline (Ti), the temperature of the mixture of blood with saline (T),and the infusion rate (Qi) provided that blood and saline are properly mixed. Takingthe differences in specific heat and density of the blood and the saline into accountand express T and Ti as the deviation with respect to the blood temperature (Tb) theoriginal blood flow can be calculated using equation 5.6:

Qth = 1.08[Ti

T− 1

]Qi + Qi (7.1)


7.3.1 Patient selection

Thirty-five patients referred for elective percutaneous coronary intervention or intra-coronary physiologic measurements were studied. These patients were selectivelychosen from our regular population on the basis of the following criteria: in case ofa stenotic artery, there should also be a segment of at least 3 cm without major sidebranches proximal to the index stenosis for reasons to be explained later. In caseof contralateral normal or almost normal coronary arteries, a segment with a length

Coronary flow measurement by continuous infusion thermodilution: Human study 73

of at least 3 cm without major side branches had to be present. Patients with verytortuous coronary arteries and chronic total occlusions were excluded. There were nofurther exclusion criteria. The study was approved by the institutional review board,and informed consent was obtained from all patients.

7.3.2 Cardiac catheterisation and experimental protocol

A 6F arterial and a 5F venous sheath were introduced into one femoral arteryand vein, respectively. After administration of 5000 IU of heparin, a guidingcatheter was advanced into the coronary ostium. Intracoronary nitroglycerin (200µg) was administered, and a reference coronary angiogram was made. Next, theinstrumentation was performed as depicted in figures 7.1 and 7.2. A commerciallyavailable sensor-tipped pressure/temperature guidewire (PressureWire-5, Radi Med-ical Systems, Uppsala, Sweden) was advanced into the distal part of the coronaryartery. Intra venous adenosine was administered (140 µg/(kg · min) ), resulting inrapid achievement of hyperemia, and fractional flow reserve (FFR) was determined.After FFR was determined the adenosine was stopped. Consequently an over-the-wire infusion catheter (Tracker-18 Soft Stream Sidehole Catheter, Boston Scientific,Minneapolis, MN) was advanced over the guidewire into the coronary artery. Thisinfusion catheter has 3 sideholes in the distal 3 cm of the catheter.

In case of a stenotic artery, the infusion catheter was placed proximal to thestenosis with all sideholes positioned distally from major side branches, if present.In the normal or almost normal coronary arteries, the infusion catheter was placedin the 3 cm long segment without major side branches. The position of thepressure/temperature sensor was chosen to be 3-6 cm distal to the tip of the infusioncatheter based in the experiments performed in-vitro and in animals (chapter 5 and6).

After instrumentation, hyperemia was induced again and the blood temperature(Tb) was measured by the pressure/temperature sensor. Next, saline at roomtemperature was infused through the infusion catheter at a constant rate (Qi),using a dedicated infusion pump (Angiomat 6000, Liebel-Flarsheim, Germany).During steady state continuous infusion of saline, the temperature of the bloodafter adequate mixing with the infused saline (T) was measured. Thereafter, thepressure/temperature sensor was pulled back into the infusion catheter, and thetemperature of the saline (Ti) at the location of the most proximal side hole wasmeasured. Volumetric blood flow in the respective coronary artery was calculatedusing equation 6.1.

Next, adenosine was stopped and restarted after 3 minutes, whereafter thecomplete measurement cycle was repeated for assessing reproducibility. In patientswith no significant epicardial stenosis (FFR>0.75), no intervention was performed. Inthe other patients with a significant coronary artery stenosis (FFR<0.75), a stent wasplaced and the thermodilution-based flow measurements were performed both beforeand after stenting of the functional significant lesion. At the end of the procedure,after removing the infusion catheter, angiograms from 2 directions were repeated andalso FFR measurements were always repeated.

74 Chapter 7

Figure 7.1: Experimental set-up during catheterisation. The set-up was comparableto the set-up used in the animal experiments (figure 6.1) except for the perivascularflow probe and the inflatable perivascular occluder. An commercially availableinfusion catheter, with three side holes in its most distal part and an open end hole,was advanced into the coronary artery over a 0.014” pressure/temperature sensor-tipped guidewire through a Y-connector (Y1). The infusion catheter was positionedwith its side holes in a segment without major side branches. The infusion catheterwas then connected to an infusion pump by a second Y-connector (Y2), enablingcontinuous infusion of saline at room temperature. The sensor-tipped guidewire wasconnected to the interface (RADI analyzer) as routinely done in coronary pressuremeasurements and distal pressure (Pd) and temperature (T) were displayed on theinterface. Also the aortic pressure (Pa), measured at the tip of the guiding catheter,was recorded by a regular pressure transducer and displayed on the interface (seefigure 7.3).


Figure 7.2: Instrumentation of the right coronary artery of a 57-year-old man. Thesensor-tipped guide wire is advanced into the coronary artery. Consequently theinfusion catheter is advanced over the guide wire. Its tip is indicated by the middlearrow. The sensor of the guide wire is indicated by the bottom arrow.

In figure 7.3 an example of a temperature tracing is shown. During all temperaturemeasurements, aortic and distal coronary pressures were continuously recorded anddisplayed. In the patients undergoing stenting, also coronary occlusion pressure (Pw)was recorded during a one-minute balloon inflation to enable separate measurementof coronary and myocardial fractional flow reserve (Pijls et al., 1993, 1995b), as willbe explained later.

7.3.3 Measurement procedure

Reproducibility

To assess reproducibility, measurements were repeated with an interval of 3 minutesin all patients. In 7 of the patients, measurements were also performed in a normal oralmost normal contralateral artery. So, a total of 42 coronary arteries in 35 patientswere studied in this way.

Variation of infusion rate

To determine the effect of a different infusion rate (Qi), in eleven patients themeasurement of absolute coronary blood flow was repeated using a different infusionrate. The infusion rate was varied between 10 and 25 ml/min. No infusion ratesbelow 10 ml/min were used to avoid unfavorable signal-to-noise ratios, and infusionrates above 25 ml/min were not used to avoid excessive cooling of the myocardiumwith possible adverse effects on the conduction system.

76 Chapter 7

Figure 7.3: Example of thermodilution measurement in a patient. On the left, bloodtemperature at steady state hyperemia (Tb) is set to zero. Next, infusion of saline isstarted and temperature T is recorded. Next, the sensor is pulled back to the tip ofthe infusion catheter to measure Ti. Using equation 7.1 coronary blood flow can becalculated. Coronary blood flow is calculated comparable to figure 6.2.


Variation of sensor position

To investigate the influence of different sensor positions within the coronary artery oncalculated flow values, in ten other patients the measurements were repeated usinga different position of the pressure/temperature sensor. Sensor position was variedbetween 3 cm and 6 cm distal to the tip of the infusion catheter.

Indirect quantitative validation

Because direct validation of this technique for measuring absolute coronary flow is notpossible in a closed-chest human model, indirect validation was performed in thosefourteen patients undergoing stenting in an indirect way. First, flow was measuredby the thermodilution technique before the intervention and called Qth,pre. Next, astent was placed and coronary occlusion pressure Pw was measured during ballooninflation. From the hyperemic coronary pressure measurement before stenting,during occlusion, and after stenting, coronary fractional flow reserve (FFRcor) can bedetermined before and after stenting. FFRcor expresses maximal achievable coronaryblood flow in the presence of a stenosis as the ratio to the coronary blood flow if theartery would be completely normal (Pijls et al., 1993, 1995b). Therefore, the ratiobetween FFRcor after and before stenting indicates increase of maximum coronaryblood flow achieved by stenting (Pijls et al., 1993, 1995b).

After stenting blood flow was measured again using the new thermodilutiontechnique (Qth,post). The ratio between Qth,post and Qth,pre should reflect the sameincrease in coronary flow as determined by the pressure based FFRcor measurements.Indirect evidence of the quality of the new method is obtained in this way.


Reproducibility, the influence of a different infusion rate of saline, the influence ofsensor position, and comparisons of the FFRcor and Qth ratios were evaluated bylinear regression analysis and with Bland-Altman plots. The best linear fit was foundwith linear regression analysis (Matlab, The Mathworks Inc., Natick, MA, USA).Hemodynamic data are given as mean ± standard deviation.

7.4 Results

7.4.1 Baseline characteristics and procedural results

The baseline characteristics of the patients are presented in table 7.1. As might beexpected from the inclusion criteria, the RCA was over-represented. Out of the 35patients, 14 had a hemodynamic significant stenosis (FFR=0.67±0.17) in the rightcoronary artery (n=10), left circumflex artery (n=1), and left anterior descendingartery (n=3), respectively. In all of these patients, successful stenting was performed.The other patients had no significant stenosis and the measurements were performed

78 Chapter 7

in the index coronary artery in all of them and in an apparently normal contralateralartery in 7 of them.

In two of these patients, early in the study, transient atrioventricular conductionabnormalities occurred when distal blood temperature became below 33◦C. In allsubsequent experiments, therefore, infusion rate of saline was chosen in such a waythat the distal temperature was between 0.5 and 2.0◦C below blood temperature(infusion rate 10-25 ml/min, depending on the size of the artery and the severity ofthe stenosis). No further complications occurred in any of the patients.

Once the instrumentation was achieved, which lasted approximately ten minutesper patient, the time needed to perform thermodilution measurements itself was lessthan three minutes per measurement in all patients. Absolute hyperemic coronaryblood flow in the stenotic coronary arteries was 96±38 ml/min and increasedto 141±50 ml/min after PCI. This corresponded with FFRmyo of 0.67±0.17 and0.89±0.04 respectively and an increase of FFRcor of 0.58±0.21 to 0.87±0.05,respectively. In those coronary arteries without significant stenosis and fourteen notstented on the basis of FFR, absolute blood flow was 141±55 ml/min. In the patientsundergoing stenting, follow-up treatment was according to local routine.

7.4.2 Flow measurements

Reproducibility

As mentioned, reproducibility was tested in 42 coronary arteries in 35 patients.The agreement between the first and the second measurement is presented infigure 7.4(a). An excellent reproducibility was obtained (Qth,2 = 1.0 × Qth,1 + 0.9ml/min, R2=0.97).

Infusion rate

In eleven arteries, the measurement was performed at two different infusion rates (1025 ml/min). Infusion rate was chosen such that distal coronary blood temperature atsteady-state infusion was in the range between 0.5 to 2.0◦C below blood temperature.Our experiments rapidly taught us that in a large, almost normal coronary arteryinfusion rate should be between 15 – 25 ml/min and in a stenotic artery 10 – 20ml/min. If infusion rate at the initial measurement was chosen too low (resulting ina distal temperature during infusion less than 0.5◦C below blood temperature), theinfusion rate was increased. In all experiments, we started with the lower infusionrate and increased this rate by a factor 1,5 – 2. A good agreement was obtained(Qth,high = 1.1 × Qth,low + 9.0 ml/min, R2=0.87). The results of this variation ininfusion rate are also presented in figure 7.4(b).

Position

In ten coronary arteries, measurements were repeated using a different position ofthe pressure/temperature sensor. This position was chosen between 3 cm and 6 cmdistal to the tip from the infusion catheter. If the sensor position was chosen less


Table 7.1: Baseline patient characteristics, FFR, and measured blood flowNo. of patients 35No. of coronary arteries 42Male/female 23/12Age (yrs±SD) 55±11LAD/LCX/RCA 10/1/31Current smoking 11Diabetes 4Hypertension 14Dyslipidemia 18Prior infaction in target area 5

Arteries with FFR>0.75 (n=28)FFRmyo 0.86±0.06Qth (ml/min) 141±55

Arteries with FFR<0.75 (n=14)FFRmyobefore stenting 0.67±0.17FFRmyoafter stenting 0.89±0.04FFRcorbefore stenting 0.58±0.21FFRcorafter stenting 0.87±0.05Qth before stenting (ml/min) 96±38Qth after stenting (ml/min) 141±50

FFR = fractional flow reserve; FFRcor = coronary fractional flowreserve; FFRmyo = myocardial fractional flow reserve; LAD = leftanterior descending artery; LCX = left circumflex; RCA = rightcoronary artery; Qth = volumetric coronary blood flow calculated bythermodilution (equation 7.1); SD = standard deviation.

80 Chapter 7

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qth,1

[ml/min]

Qth

,2 [m

l/min

]

Reproducibility

(a)

y =1x+0.93R2 =0.97

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,1

and Qth,2

[ml/min]

Rel

. diff

. Qth

,1 a

nd Q

th,2

[%]

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qth,low

[ml/min]

Qth

,hig

h [ml/m

in]

High vs Low infusion rate

(b)

y =1.1x+9.4R2 =0.87

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,low

and Qth,high

[ml/min]

Rel

. diff

. Qth

,low a

nd Q

th,h

igh [%

]

0 50 100 150 200 250 3000

50

100

150

200

250

300

Qth,prox

[ml/min]

Qth

,dis

t [ml/m

in]

Proximal vs Distal position

(c)

y =0.99x+3.4R2 =0.94

0 50 100 150 200 250 300−100

−75

−50

−25

0

25

50

75

100

Average Qth,prox

and Qth,dist

[ml/min]

Rel

. diff

. Qth

,pro

x and

Qth

,dis

t [%]

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

FFRcor,pre

/FFRcor,post

[−]

Qth

,pre

/Qth

,pos

t [−]

Relation FFR and Qth

(d)

y=0.75x+0.2R2 =0.84

0 0.2 0.4 0.6 0.8 1−100

−75

−50

−25

0

25

50

75

100

Average ratios FFRcor

and Qth

[−]

Rel

. diff

. rat

ios

FF

R cor a

nd Q

th [%

]

Figure 7.4: Results for the human experiments. Second versus first measurement forall experiments, represented by linear regression and corresponding Bland Altmandiagram (a). High versus low infusion rate (b). Distal versus proximal location ofsensor (c). Indirect validation of the method (d).


than 3 cm from the tip, often undulations in the temperature tracing occurred, mostlikely due to incomplete mixing of the saline and blood and making it impossible tointerpret the right value for T. A sensor position more than 6 cm distal to the tip ofthe infusion catheter was deemed to be inappropriate because previous laboratorytesting showed that in such a case, loss of indicator (cold saline) into the wall is notnegligible anymore and flow might be overestimated. The results of measurementsat two different sensor positions are also shown in figure 7.4(c). Again, an excellentagreement was present and it is evident that the sensor position within this range is ofno significant influence on the results of the measurements (Qth,dist = 0.99 × Qth,prox+ 3 ml/min, R2=0.94).

Indirect quantitative validation

In fourteen patients, the in-duplo measurements were performed both before andafter stenting of a coronary artery stenosis. The ratio of the flow values measured bythe thermodilution technique before and after stenting (Qth,pre , and Qth,post) wascompared to the ratio of FFRcor values before and after stenting, as explained inthe methods section. A good agreement was found between Qth,pre /Qth,post andFFRcor,pre /FFRcor,post as shown in figure 7.4(d), corroborating the validity of ourmethodology .

7.5 Discussion

In this study we showed that the thermodilution technique for measuring volumetricblood flow in selective coronary arteries is feasible in conscious humans. Moreover,the technique is fairly simple and straight forward to perform, reproducible, andsafe. Compared to the coronary sinus measurements performed by Ganz et al.(1971) in the seventies the variability is smaller, instrumentation is less complicatedand measurements can be performed easily and quickly with standard equipmentduring coronary interventions using the diagnostic pressure/temperature guidewire,the corresponding interface (RADI Analyzer) and a suitable infusion catheter.

In accordance to our in-vitro and animal experiments (chapter 5 and 6) theinvestigated sensor positions did not affect the measurements indicating that ”loss”of indicator through the wall due to heating was negligible between 3 to 6 cm distalto the tip of the infusion catheter. Moreover, the accuracy of the thermodilutiontechnique was not significantly affected by different infusion rates either.

We learned that the amount of infused indicator has to be kept within a certainrange, comparable to the infusion rates used in the animal experiments (i.e. 10-25 ml/min), depending on the artery. In the pilot stage of the experiments higherinfusion rates of saline caused excessive cooling of the myocardium, resulting inconduction disturbances, which occurred twice in our series when distal temperaturefell below 33◦C. Therefore, we chose the infusion rate in such a way that distaltemperature decreased to a value 0.5 to 2.0◦C below blood temperature at steadystate infusion.

82 Chapter 7

Being able to determine absolute coronary blood flow in the catheterisationlaboratory in a fairly simple and straightforward way has several advantages.First, it enables us to express coronary blood flow in selective coronary arteriesquantitatively during cardiac catheterisation. This has not been possible so far.Second, when performing PCI, also myocardial blood flow and collateral blood flowcan be calculated because also Pw can be measured by the same guidewire, therebyproviding separate contribution of coronary and collateral flow to myocardial bloodflow (Pijls et al., 1993). Third, because also distal coronary pressure is measuredsimultaneously by the same sensor, myocardial resistance of a specific myocardialterritory can be easily calculated: myocardial resistance equals myocardial perfusionpressure divided by myocardial blood flow (Fearon et al., 2004; Aarnoudse et al.,2004a). Quantification of myocardial resistance will give a better insight intomicrocirculatory dysfunction, and will make it possible to selectively study themicrocirculation, whereas the fractional flow reserve, simultaneously measured bythe same equipment, selectively interrogates the epicardial artery. Further studies inparticular groups of patients in whom microcirculatory function is affected, will bepossible in this way, such as in patients after myocardial stem cell therapy, or duringinvasive follow-up of patients after myocardial infarction or heart transplantation.

7.5.1 Limitations

Instrumentation for the application of the thermodilution technique in a coronaryartery is somewhat more complicated than in regular diagnostic or interventional in-tracoronary procedures. In our studies, these extra steps required only approximatelyten minutes. Once the instrumentation was accomplished, the measurements couldbe performed quickly and easily. Nevertheless, the need for a second Y-connectorand the infusion pump will most likely limit the use of this method to investigationalstudies and specific groups of patients.

Despite no direct validation was possible in this closed-chest human experimentan indirect validation was performed based on the well validated hyperemic pressuremeasurements and accompanying values for FFR. It should be noted that coronaryflow, as determined by the thermodilution technique, might not be equal to myocar-dial flow. In normal coronary arteries, it can be hypothesized that myocardial bloodflow does equal coronary blood flow. In the presence of a stenosis, however, thecollateral component plays an increasingly important role with increasing stenosisseverity (Fearon et al., 2004; Aarnoudse et al., 2004a). Fortunately, if stenting isperformed coronary wedge pressure (Pw) can be measured. This enables calculationof FFRmyo / FFRcor. Because FFRmyo and FFRcor represent the mutual relationbetween maximum myocardial and coronary flow, measurement of absolute coronaryflow also enables calculation of absolute myocardial blood flow in that situation andthereby also absolute collateral flow as the difference of the former two (Fearonet al., 2004; Aarnoudse et al., 2004a). Also collateral blood flow can be calculatedquantitatively in that situation by subtracting coronary blood flow from myocardialblood flow.

Another technical difficulty is given by the infusion catheter itself. In this human


study, we used a commercially available infusion catheter (Tracker 18, soft stream20 side hole catheter). This catheter was tested extensively in the in-vitro set-up asdiscussed in chapter 5 (catheter A). The fact that the side holes were distributed overa length of 3 cm limited us to study only patients in whom no major side branchesoriginated from the segment scheduled for measurement. Hence, in the presence ofmajor side branches, indicator might be lost before complete mixing occurred andcoronary flow would be overestimated. Besides the limited anatomic applicability,the results for this infusion catheter obtained in the experiments in chapter 5 showedan underestimation of calculated flow with increasing coronary flow values, probablydue to incomplete mixing. Besides the fact that mixing will be positively affected bythe application in-vivo , the coronary flow values were ±150 ml/min on average.In this range, even the agreement between Qth and actual coronary flow Q wasreasonable for this catheter with an open end hole as found in the in-vitro experiments(see section 5.3). Especially to overcome the limited anatomic applicability, a specificinfusion catheter was designed in the meantime with 4 small laser punched side holesover the last 0.5 cm (based on the experimental findings of catheter B in chapter 5).

Being able to measure coronary and myocardial flow, it would also be possible tocalculate myocardial resistance because mean distal coronary pressure is measuredsimultaneously by the same sensor; mean myocardial resistance equals meanmyocardial perfusion pressure (measured by the sensor) divided by mean myocardialblood flow (Aarnoudse et al., 2004a; Bol et al., 1993).

Although absolute coronary blood flow has been considered as the ultimategoal of coronary physiologists for decades, the concept itself has several limitationsfor practical purposes as already mentioned in 6.5.1. Moreover, to study if, e.g.,microcirculatory function is affected or improved in patients after myocardial stemcell therapy, myocardial infarction, or heart transplantation, it is not the absolutevalue of the flow or microvascular resistance that is of importance but the changeof the parameter over time. Since reproducibility was good in both the animal andpatient study on the short term, it might be expected that the changes of flow orresistance over time can be detected.


Determination of absolute blood flow in selective coronary arteries quantitatively inml/min in a fairly simple and straightforward way during cardiac catheterisation inhumans, has not been possible so far and has several advantages: Not only coronaryflow in itself is measured, but because the guidewire also measures distal coronarypressure and fractional flow reserve, values of flow can be directly related to itsnormal maximum value and therefore better be interpreted.

Quantification of (changes in) myocardial resistance makes it possible to selec-tively study the microcirculation, whereas the fractional flow reserve, simultaneouslymeasured by the same equipment, selectively interrogates the epicardial artery.Further studies in particular groups of patients in whom microcirculatory functionis affected, will be possible in this way, such as in patients after myocardial stem celltherapy, myocardial infarction, or heart transplantation.

84 Chapter 7

So, we believe that this novel methodology for direct volumetric coronaryflow measurement can be a valuable tool in the clinical research of the coronarycirculation.

7.6 Conclusion

We described a method to measure coronary blood flow in selective coronary arteriesduring cardiac catheterisation in conscious humans, using a thermodilution techniquewith continuous infusion of saline at room temperature. As an indirect measure foragreement, improvement in flow determined by the thermodilution technique beforeand after stenting corresponded well with values for flow improvement based onpressure measurements. Together with distal coronary pressure measurement, mea-sured by the same sensor simultaneously, it would be possible to calculate absoluteresistance of the coronary artery and coronary microcirculation. This methodology,therefore, constitutes the first direct volumetric blood flow measurement in selecthuman coronary arteries and enables quantitative assessment of microcirculatoryfunction.

Acknowledgement

The authors wish to express their gratitude to the fabulous nurses and technicians ofthe catheterisation laboratory of the Catharina Hospital.

Chapter 8

Biomechanical properties ofabdominal aortic aneurysmsassessed by simultaneously

measured pressure and volumechanges in humans

Abdominal aortic aneurysms (AAA) are at risk of rupture when the internal load (blood pressure)exceeds the strength of the wall of the aneurysm. Generally, the maximal diameter of the aneurysmis used as a predictor of rupture. Biomechanical properties may be a better predictor for rupturethan the maximal diameter. Compliance and distensibility are mechanical properties which canbe determined by measuring the pressure-volume relationship of the aneurysm. The objectiveof this study is to determine the compliance and distensibility of the AAA of 10 patients bysimultaneous instantaneous pressure and volume measurements. In all patients we found a stronglinear relation between the pressure and volume data. Average compliance was 0.31±0.15 ml/kPawith accompanying estimates for Young’s moduli 9.0±2.5 MPa. Distensibility was 1.8±0.7 10−3

kPa−1. For the first time compliance and distensibility of the wall of the aneurysm is determinedin-vivo in conscious humans. MRI-based compliance monitoring is potentially useful for onlinerupture risk assessment and follow-up of aneurysms in combination with accurate (non-invasive)systolic and diastolic pressure measurements.

The contents of this chapter are based on: M. van ’t Veer, J.Buth, M.A.G. Merkx, P.A.L.Tonino, N.H.J. Pijls, H. van den Bosch, and F.N. van de Vosse. Biomechanical propertiesof abdominal aortic aneurysms assessed by simultaneously measured pressure and volumechanges in humans. J Vasc Surg, in press.

85

86 Chapter 8

8.1 Introduction

The rupture rate associated with AAA smaller than 5.5 cm has been shown to be1% per year (The-UK-Small-Aneurysm-Trial-Participants, 2002). However, it has alsobeen reported that 5-10% of the ruptured aneurysms are below this diameter criterion(Nicholls et al., 1998). Conversely, many AAAs larger than this cut-off value will notrupture within the patient’s life time (Conway et al., 2001; Powell and Brown, 2001).In an attempt to develop methods to better predict the risk of rupture, biomechanicalparameters have been investigated. Compliance is a biomechanical characteristic ofthe vessel which is defined as volume change resulting from change in intra-vascularpressure (equation 8.1). A vessel is defined compliant or distensible when a smallchange in pressure translates into a large volume change. For several diseases thatare characterized by degradation processes of the vessel wall, for example Marfan’sdisease or atherosclerosis are associated with changes in compliance (Lalande et al.,2002; Richter and Mittermayer, 1984). Moreover, it has been demonstrated thatthere is a tendency towards an increase in distensibility in patients who experiencerupture (Wilson et al., 2003). These data are supported by additional observations oflower tensile strength in aneurysms operated for rupture compared to non rupturedaneurysms (Martino et al., 2006). These observations emphasize the need for amethod to monitor the biomechanical status of the aneurysmal wall during follow-up in-vivo, in addition to regular assessment of the maximal diameter.

As a surrogate for volume change, change in cross sectional area or diameterof the vessel is frequently determined and compliance is calculated per unit length(Lalande et al., 2002; Lang et al., 1994). Also pressure-strain elastic modulus (Ep)is often used as a biomechanical property of the wall (Lang et al., 1994; Petersonet al., 1960). However, these assessments lack information of the target vessel asa whole and focus on changes in maximal diameter or cross sectional area only(Wilson et al., 2003; Long et al., 2005; Ganten et al., 2008; Vorp et al., 1996). Inthese studies the blood pressure is usually assessed by brachial cuff measurementor by plethysmography. These techniques measure pressure remote from the areaof interest and often relate to the pressure in the ascending thoracic rather than inthe abdominal aorta (van Bortel et al., 2001). It is not precisely known how suchperipheral pressures approximate the true intra-aneurysmal pressure and whetherreflections of pressure waves at the aneurysm and the aortic bifurcation may influencethe measurement.

In this study, we determined volume changes of the AAA through the cardiaccycle simultaneously with invasively measured intra-aortic pressures within the AAA.Pressure-volume relations and subsequently distensibility and compliance of theabdominal aortic aneurysm were determined from dynamic MRI and intra-arterialpressure data. To our knowledge this is the first study in which compliance anddistensibility of an AAA were determined from simultaneously measured intra-aneurysmal pressure and volume change in humans. Whether brachial pressures canbe used to replace intra-aneurysmal measurements was studied as a secondary goal.

Biomechanical properties of AAA 87

8.2 Methods


Ten patients who had an indication for repair of their infrarenal aortic aneurysmeither by endovascular stentgrafting or open surgical treatment were included inthe study. Patients with contraindications for magnetic resonance imaging (MRI),cardiac arrhythmia, and obstructed iliac arteries were excluded from this assessment.The study was approved by the institutional review board of the Catharina HospitalEindhoven, the Netherlands, and written informed consent was obtained from allparticipating patients prior to enrolment in the study.

8.2.2 Magnetic Resonance Imaging

Magnetic resonance imaging was performed on a Philips Gyroscan Intera 1.5 Tesla MRscanner (Rel. 10.4) (Philips Medical Systems, Best, The Netherlands). The scanningarea was defined as the area proximal to the renal arteries until just beyond thebifurcation. Three-dimensional balanced turbo field echo (3D B-TFE) images wereacquired for 50 slices with an overlap of 3 mm. Images were acquired for 15 cardiacphases (SENSE cardiac coil, TE/TR=2.14/4.28 ms, flip angle 50 degrees, FOV=300mm, voxel dimensions 1.2×1.2×6 mm, slice gap 0 mm, no breath-holding, noncontrast enhanced). Images were stored and analyzed offline for volume changesand calculating biomechanical properties.

8.2.3 Pressure measurement

Pressure measurement was performed by a multi-hole 6F fluid filled diagnostic pigtailcatheter (Cordis, Johnson & Johnson, Miami, FL, USA) connected to a disposablepressure transducer (Becton, Dickinson and Co., Franklin Lakes, NJ, USA). Thecatheter was introduced via one of the femoral arteries. Prior to the introductionof the catheter 5000 IU of heparin were administered. The tip of the catheter wasplaced in the AAA halfway between the renal arteries and the iliac bifurcation. Smallinjections of contrast dye were used to verify the correct position fluoroscopically.Subsequently the catheter was connected to the pressure transducer. An MRIcompatible flush system was used to prevent clotting of blood in the catheter duringimage acquisition. The pressure transducer was connected by an extended electriccable to a pressure recording interface (RADI Analyzer, RADI Medical Systems,Uppsala, Sweden) outside the MRI-scanner room. Pressure was calibrated to openair. The pressure measurement set-up was tested extensively in-vitro before usageand errors were less than 5%. Magnetic resonance images were acquired as describedbelow and during the imaging sequence the intra-aneurysmal pressure was recordedcontinuously. Pressure measurements were stored and analyzed offline.

Immediately before and after the imaging sequence the brachial artery bloodpressure was measured with an automated MRI-compatible sphygmomanometer(Magnitude 3150 MRI, Invivo Co., Orlando, FL, USA) determined simultaneously

88 Chapter 8

with the continuously measured invasive blood pressure. These measurements wererepeated three times for each patient. After the MRI protocol was completed, thepigtail catheter and the arterial sheath were removed under fluoroscopic guidanceand a pressure bandage was applied.

The pressure corresponding to each cardiac phase was determined by perphase averaging of the directly measured pressure. Heart beats that deviatedmore than 20% compared to the heart rate that was used as an input for theimaging sequence, were not included in the analysis of image reconstructionand pressure measurements. Averaged invasively measured systolic and diastolicpressures were compared to the simultaneously measured non-invasive brachialartery blood pressures and represented in Bland Altman plots. A paired signed ranktest was performed to compare systolic and diastolic blood pressure values betweenthe invasive and non-invasive technique.

8.2.4 Detection of volume changes

Image post-processing was performed on slices ranging from just distal to the lowestrenal artery until just proximal to the aortic bifurcation. Prototype software (PhilipsMedical Systems, Best, The Netherlands) based on an existing cardiac package wasused to detect volume changes of the aneurysm (Hautvast et al., 2006). The semi-automated software required a manually drawn initial contour of the cross sectionof the aneurysmal wall in each slice. Subsequently, the contour was propagatedautomatically through the cardiac phases. The catheter, located in the lumen, didnot influence the propagation. The resulting contours of the different slices weremultiplied by the slice thickness, while taking overlap into account, to obtain volumesand volume changes through the 15 cardiac phases. In pre-study testing inter-uservariability of for determination of volume canges using this technique was found tobe small: 2.4±4.7%.

8.2.5 Biomechanical properties of the AAA

For each individual patient, volume was plotted against pressure to obtain pressurevolume loops (figure 8.1). When assuming small strains, the volume and pressuredata can be linearized as a first approximation for an aortic wall model. The bestlinear fit was found with simple linear regression analysis (Matlab, The MathworksInc., Natick, MA, USA). Consequently, the slope of the best fit represents the value forthe compliance C of the AAA:

C =∆V

∆P(8.1)

For estimation of the Young’s modulus we assumed a thin walled cylinder withaxial movement restraint at both ends. The radius was chosen such that the diastolicvolume of the aneurysm equalled that of a cylinder with the same length. From


Laplace’s law the Young’s modulus can be derived:

E =1

7.52V0Rav

h

(1− ν2)C

(8.2)

where Rav represents the radius of the cylinder (with diastolic volume V0), h thewall thickness, ν Poisson’s ratio, and C the compliance (in ml/kPa). When we rewriteequation 8.2, it should be clear that there is a relation with the Young’s modulus (E)we defined and the pressure-strain elastic modulus (Ep) as defined by Peterson et al.(1960) in terms of pressure differences and differences in diastolic (D0) and maximal(Dmax) diameter:

E =Rav(1− ν2)

h

∆P · V0

(Vmax − V0)is related to: Ep =

∆P ·D0

(Dmax −D0)

To compare biomechanical properties within and between patients it should beindependent of the initial volume. Compliance, however, increases for larger volumes.Distensibility (D) is a biomechanical property that takes the initial volume of the AAAinto account:

D =1V0

∆V

∆P(8.3)

where V0 equals the diastolic volume of the AAA. Distensibility is expressed inPa−1. For these biomechanical parameters the pressures are expressed in Pa (100mmHg corresponds to 13.33 kPa).

8.3 Results

8.3.1 Baseline characteristics and clinical results

All patients participating in the study were male. The average age (±SD) was73.6±6.4 years. The average maximal diameter of the aneurysm was 5.8±0.6 cm.Three out of ten patients were using drugs for their hypertension. Instrumentationwas uncomplicated in all patients and none of the patients experienced any adverseevents. Eight patients received an endovascular stentgraft and two patients under-went open surgery. Time from MRI until surgery or stentgraft placement was 22±12days.

8.3.2 Aneurysmal volume change and mechanical properties

Volume change propagated over the cardiac cycle resulted in an average of 3.0±1.1ml for all patients. Diastolic and systolic volume data for each patient are depicted intable 8.1.

90 Chapter 8

Table 8.1: Aneurysmal volume dataPatient nr. Diastolic volume Systolic volume

[ml] [ml]1 129 1332 211 2153 182 1854 110 1115 172 1756 230 2347 154 1568 207 2119 193 19610 126 127

Compliance was determined as the slope of the best linear fit (figure 8.1). Valuesfor the compliance of each patient as well as a measure for the goodness of the fit(i.e. correlation coefficient) are shown in table 8.2. Taking the diastolic volumes intoaccount, distensibility (D) was calculated using equation 8.3. These results are alsopresented in table 8.2. Good linear fits were found for all but two patients. Patient 6and 10 showed R2 of 0.62 and 0.27 respectively. These low values for the correlationcoefficients are the result of unreliable volume measurements from the cardiac phases8 till 15 (declining part) in both patients resulting from poor image quality due toirregular heart beat and patient movement. Uncertainty remains whether the AAAsof these patients show a linear relation between pressure and volume. In retrospect,these two patients did not satisfy the inclusion criteria completely.

Since MRI fails in detecting wall thickness, a value of 2 mm was assumed forcalculation of the Young’s moduli (Arko et al., 2007). Moreover, the wall was assumedto be incompressible (ν = 0.5). Estimations for the Young’s moduli ranged from 5.5to 12.9 MPa. Values for each patient are presented in table 8.2.

8.3.3 Invasive and non-invasive blood pressure

The comparison of the invasively and non-invasively measured blood pressuremeasurements is represented in figure 8.2 using a Bland Altman plot. On averagethe systolic blood pressure measured with the brachial cuff underestimates theintra-aneurysmal pressure measured by the pigtail catheter by 5% (figure 8.2 A,(P<0.001)). On average the diastolic blood pressures measured with the brachialcuff overestimates the intra-aneurysmal pressure measured by the pigtail catheter by12% (figure 8.2 B, (P<0.001)).


Table 8.2: Mechanical properties of the aneurysmsPatient nr. D C R2 E

[10−3 kPa−1] [ml/kPa] [MPa]]1 2.2 0.29 0.97 7.02 3.0 0.64 0.92 5.53 1.9 0.35 0.95 9.74 1.8 0.20 0.83 7.95 2.2 0.38 0.95 6.96 1.2 0.28 0.62 9.87 1.8 0.28 0.91 7.18 1.8 0.36 0.96 11.19 1.2 0.23 0.93 12.0

10 0.5 0.07 0.27 12.9D is distensibility,C is compliance, R2 is correlationcoefficient, and E is the Young’s modulus

0 5 10 15206

208

210

212

Number of phase

Vol

ume

of A

AA

in m

l

Volume

0 5 10 155

15

25

Number of phase

Pressure

Pre

ssur

e in

AA

A in

kP

a

5 15 25206

208

210

212

Pressure in AAA in kPa

Vol

ume

of A

AA

in m

l

Pressure−Volume

datapointsPV−relationbest fit

y = 0.36 + 203.8R2=0.96

Figure 8.1: Simultaneous pressure and volume registrations of the aneurysm. Thevolume of the aneurysm is determined for fifteen cardiac phases from the MRI images(left panel). Average values for the pressure within the aneurysm are determinedfor the same cardiac phases (mid panel). A strong linear relation between the twovariables is found (right panel). The slope of the best linear fit reflects the complianceof the aneurysm.

92 Chapter 8

100 110 120 130 140 150 160−50

−25

0

25

50

average of methods

Rel

diff

[non

inv

BP

and

inv

BP

] [%

]

Bland Altman relative differences systolic BP measurements

60 65 70 75 80 85 90−50

−25

0

25

50

average of methods

Rel

diff

[non

inv

BP

and

inv

BP

] [%

]

Bland Altman relative differences diastolic BP measurements

Figure 8.2: Bland-Altman plots for the systolic (top) and diastolic (top) bloodpressure measurement (in mmHg). The relative difference between the distantnon-invasive brachial cuff measurement and the local intra-aortic invasive bloodpressure measurement is plotted against the average of these two techniques. Thethick continuous lines depict the average relative difference between the techniqueswhereas the dashed lines depict the variation (1.96 times the standard deviation eachside) of the relative difference. Systolic blood pressure is underestimated by 5%whereas diastolic pressure is overestimated by 12% by the brachial cuff compared tothe invasively measured pressure.


8.4 Discussion

Distensibility and compliance of abdominal aortic aneurysms were determined in-vivofor 10 patients by simultaneously measuring pressure and volume changes of the AAA.A strong linear relation exists between pressure and volume. To our knowledge thisis the first study to calculate AAA distensibility and compliance in-vivo from volumechanges and simultaneously invasively measured pressure within the AAA.

In clinical practice the maximal aortic diameter is the main determinant fordecision making with respect to treatment of AAA. By convention all aneurysms witha diameter less than 5.5 cm are defined small. Results of two randomized trialshave shown equal efficacy of early intervention compared to watchful waiting untilaneurysm growth or symptoms occur (Lederle et al., 2002; The-UK-Small-Aneurysm-Trial-Participants, 1998). This does not resolve the problem that small aneurysmsalso may rupture. Of all ruptured aneurysms 5-10% has a maximal diameter smallerthan 5.5 cm (Nicholls et al., 1998). It seems plausible that other relevant factorswill be associated to the risk of rupture than maximal aneurysmal diameter alone.The conventional diameter criterion for active treatment disregards the importance offactors that determine the wall strength. Distensibility is a complementary parameterto the maximum aneurysmal diameter that is associated with an increased risk ofrupture as was suggested by Wilson et al. 2003. These authors, using an echotracking technique to assess pulsatile diameter change of the aneurysm, demonstratedthat there is a tendency towards increased values of distensibility in patients whoexperienced rupture of their AAA. The increased distensibility indicates that the vesselwall is less stiff and may be more prone to rupture. These findings are supported bydata from tensile tests on excised wall segments of electively operated AAAs andruptured AAAs (Martino et al., 2006). In these ex-vivo studies it was observed thattissue from ruptured aneurysms was less stiff and the tensile strength was lowercompared to electively operated AAAs. It is notable that the examined tissue sampleswere all taken from the anterior wall and not necessarily from the site of rupture.These findings indicated that the entire aneurysm is affected by a similar degree ofdegeneration that results in the weakening of the wall. However, the cause of thispathologic process remains unclear.

Strictly speaking, volume changes should be determined, instead of changes incross section (or diameter) at the maximal diameter of the aneurysm, for distensibilityand compliance measurements. After all, the degenerative process affects the entireaneurysm and arterial vasculature and looking for wall characteristic that reflects theoverall condition of the aneurysm seems a more logical approach. Notwithstandingthe strong correlation with rupture risk, maximal diameter does not necessarilyindicate the site of rupture (Vorp et al., 1998). Currently on theoretical grounds,we favor the assessment of volume changes over diameter recordings. We calculateddistensibility as the slope of the best linear fit through the pressure data and thedata on volume change of the aneurysm, while taking the initial volume of the AAAinto account. We believe that biomechanical information of the wall of the entireaneurysm, rather than the portion restricted to the maximal diameter, needs to betaken into account. To compare our data to other studies distensibility or compliance

94 Chapter 8

Table 8.3: Pressure-strain elastic modulus of the aneurysmal wall based on changesin cross sectional area

Our study Othters3.2±1.1 mmHg−1 4.0±0.9 mmHg−1∗

2±0.5·10−6 Pa−1 6±5·10−6 Pa−1∗∗

∗Values based on (Vorp et al., 1996)∗∗Values based on (Ganten et al., 2008)

by changes in cross sectional area at the aneurysm, as defined by other authors, leadsto values in the same range (table 8.3).

An important finding of our study is the strong linear relation between the volumeand pressure changes for all but two patients (figure 8.1 and table 8.2). Owing to thislinear relationship it would be sufficient to obtain an accurate systolic and diastolicblood pressure value rather than a complete phasic pressure curve to calculate thedistensibility of an AAA. Corresponding volume data can then simply be scaled withthe diastolic and systolic pressure values. If pressure data can be obtained non-invasively in an accurate way even absolute values for the biomechanical propertiesof the wall of the aneurysm can be calculated and may be used in patient specificwall stress analyses. Rupture risk predictions based on patient specific wall stressanalyses require absolute values for load and (bio)mechanical properties as well asan accurate geometry of the aneurysm (de Putter et al., 2007). Risk stratificationbased on such analyses demonstrated superior results compared to predictions ofrupture on diameter information alone (Fillinger et al., 2003). We calculated Young’smoduli from the measured compliances. Our values (table 8.2) are in the sameorder of magnitude as those found by Martino et al. (2006) for the tangentialmodulus corresponding to a pressure of 100 mmHg, however, our values are 3 to4 times as high. An explanation for this difference may be that Di Martino and co-workers performed ex-vivo tensile tests on excised material. Although during thesetests efforts were made to approach the biological environment, the effects of thebench test situation on the material properties compared to the in-vivo situation areunknown. Another reason why our obtained values for the Young’s modulus werehigher compared to the ex-vivo tensile test is a different estimate of wall thickness.Since we were unable to quantify the wall thickness on MRI we assumed a value of2 mm for our calculations, whereas Di Martino reported values varying from 1.5 to 5mm. Although our assumption was based on previous wall thickness measurements inthe aneurysmal neck (Arko et al., 2007) the possibility exists that we underestimatedthis value, which can explain a difference by a factor 2 to 3 in the estimation of theYoung’s modulus. At the current time, we feel confident that the technique we usedprovides a reasonable estimate for the Young’s moduli in-vivo.

As a secondary goal we compared the invasively measured intra-aneurysmal pres-sure with non-invasively measured brachial cuff measurements to examine agreementbetween the two techniques. Our data suggested that systolic intra-aneurysmalblood pressure is slightly underestimated by brachial arm cuff measurements (by5%) whereas diastolic blood pressure is overestimated by 12% compared to the


invasively measured pressure. Several other non-invasive techniques which measureblood pressure show that the diastolic value does not change substantially throughthe arterial system (van Bortel et al., 2001; Wilson et al., 2001). However, most ofthese techniques are not validated for pressure estimation within AAAs. Therefore,pressure values based on these commonly used techniques might not be as accuratein the patient with an AAA.

Using our non-invasive blood pressure data, distensibility and compliance wouldbe overestimated if absolute values were of interest. However, for distensibility-based rupture risk stratification no absolute values but rather changes over timeappear useful (Wilson et al., 2001). Due to the linear relationship observed betweenpressure and volume change, the combination of volume change measurements byMRI and pressure assessment by brachial sphygmomanometer measurements enablesnon-invasive compliance determination in-vivo . Although no absolute values fordistensibility can be found this way, this technique can easily be applied in clinicalpractice and has the potential to be a useful method to monitor the risk of rupture asa complementary method to diameter assessment by repetitive examinations duringfollow-up.

8.4.1 Limitations

The method we used to detect volume change does not take longitudinal motionof the aneurysm into account. Out-of-plane motion of the AAA might appear asin-plane deformation, and will therefore translate into volume change, especiallywhen the aneurysm is tortuous. Since none of the aneurysms of the patients inour study were tortuous, we assumed that this effect was negligible in this study.A segmentation technique that includes tracking of vessel landmarks, provided thatthrough-plane resolution is sufficiently small, may solve this problem. A three-dimensional segmentation technique such as described by Wentz et al. (2007) mightform the basis for such software developments.

We assumed a uniform wall thickness of 2 mm to estimate the Young’s modulus.Although this is an accepted assumption in the literature on wall stress analyses(Fillinger et al., 2002; Martino et al., 2001), in reality wall thickness varies withinand among aneurysms of patients. Local wall properties would be of interest froma mechanical point of view. To calculate local wall properties, at least local wallthickness should be available. The determination of local wall thickness, however,remains a challenge for all current imaging methods.

This series in itself is not large enough to determine population estimates ofbiomechanical parameters or find correlations with known risk factors. However, webelieve that our study opens the perspective on an entirely non-invasive assessmentusing dynamic MRI and brachial pressures. Moreover, our method incorporatesdistensibility/compliance information based on the aneurysm as a whole, ratherthan at a single cross-section. This concept, when validated in a larger study, mayultimately evolve as a method of substantial clinical value for individual patientmonitoring.

96 Chapter 8

8.5 Conclusion

For the first time distensibility and compliance of the wall of the aneurysm weredetermined in humans by simultaneous intra-aneurysmal pressure and volume mea-surements. We observed a strong linear relationship between the intra-aneurysmalpressure and the volume change of the abdominal aortic aneurysm. Furthermore,brachial cuff measurements were different compared to the invasive measurementswithin the aneurysm. Therefore, non-invasive measurements can only be used toobtain a relative parameter rather than an absolute values for distensibility. MRI-based monitoring of this biomechanical parameter may become useful for onlinerupture risk assessment and follow-up of aneurysms in combination with brachialpressure measurements.

Acknowledgements

This study was financially supported by a scientific fund from the Catharina HospitalEindhoven and by the foundation ”Vrienden van het Hart”. The dedicated assistanceand patience with the MRI of Marleen Kohler and the support in the catheterisationlaboratory of Dr. Michels and Dr. Brueren, is gratefully acknowledged.

Chapter 9

General discussion andconclusions

Hemodynamic measurements throughout the cardiovascular system where the mainfocus of this work. The different projects were carried out at the departmentof Biomedical Engineering of the Eindhoven University of Technology and at thedepartments of cardiology, cardiothoracic surgery, and vascular surgery of theCatharina Hospital Eindhoven. Each project aimed to answer specific researchquestions that had arisen from clinical practice. Hemodynamic measurementsthroughout the cardiovascular system were carried out in combination with standarddiagnostic methods in order to understand pathophysiologic processes resultingfrom degenerative diseases and to clarify anatomic data by adding functionalmeasurements. Besides those specific aims, the overall goal of this work was toestablish the role of the medical engineer in a non-academic heart center.

9.1 Hemodynamic measurements

As is described in chapter 2 the cardiovascular system might be affected by atheroscle-rosis resulting in degeneration of the cardiovascular system. Often the diagnosisof pathology related to atherosclerosis is based on anatomic measurements alone.The coronary angiogram is used to detect the presence and extent of coronary arterydisease but is also used to evaluate the result of a percutaneous coronary intervention.However, it has become clear during the last two decades that trying to understandthe severity of coronary artery disease from the angiogram alone is a fundamentallyflawed approach. Hemodynamic assessment plays an increasingly important role indaily practice in the understanding of the development, progress, and treatment ofcoronary artery disease additionally to the anatomic approach.

In chapter 3 drug-eluting stents were compared to bare metal stents with respectto anatomic and physiologic properties. Stents were implanted in pairs within thesame patient to exclude the biological diversity between patients. It was found that,

97

98 Chapter 9

in contrast to the bare metal stent, the sirolimus stent maintained a normal wall shearstress value within the stented segment six months after implantation. Moreover,pressure derived fractional flow reserve at follow-up was significantly higher andthe hyperemic trans-stent gradient significantly lower for the drug eluting stent.Irrespective of the clinical consequences of the results, it can be concluded that thedrug-eluting sirolimus stent is hemodynamically superior to its bare metal counterpartwith respect to maintenance of favorable hemodynamic characteristics.

Hemodynamic measurements were also used in chapter 4 to study a pathophysio-logic situation related to valve surgery. In this chapter the aim was to determine theinfluence of the orientation of a bi-leaflet valve prosthesis on coronary perfusion. Itwas hypothesized that coronary blood flow could be influenced by a venturi effect,caused by the velocity of the ejected blood through the mechanical valve prosthesis.Vasodilatory and adrenergic drugs were used to mimic exercise conditions whichconsequently lead to maximal ejection velocities through the prosthesis. Participatingpatients receiving a mechanical bi-leaflet prosthesis were randomized to differentorientation of the prosthesis. Patients receiving bio-prosthesis were the control group.

From the results of this study it was concluded that the influence of a bi-leafletprosthesis on coronary perfusion pressure was negligible and not dependent onvalve orientation. Moreover, coronary perfusion pressures did not differ significantlybetween mechanical and biological prostheses. The practical consequence is that thecardiac surgeon does not have to bother about valve orientation and can choose theorientation which is most convenient from the implantational perspective.

The interpretation and implementation of the results as well as the design ofthe studies discussed above are specific examples that illustrate the activities of themedical engineer in a clinical environment. The cooperation with the medical doctorsto implement the obtained results in daily practice and the ability to propose well-defined research questions from a complex clinical problem, is an important task ofthe medical engineer.

Chapters 3 and 4 studied coronary physiology related to the epicardial arteries.To describe and interrogate the coronary microvasculature and to calculate absoluteresistance, however, the ability to measure absolute coronary blood flow would be de-sirable. In chapter 5 a technique for flow measurement based on continuous infusionthermodilution was developed and validated in an in-vitro set-up. Theoretically, meanabsolute coronary blood flow through a selective coronary artery can be calculatedfrom the temperature of the blood, the saline, and the mixture downstream of theinjection site, combined with a known infusion rate. Application of the technique toobtain reliable values for coronary blood flow requires the blood and the saline tobe completely mixed downstream of the site of injection. In the mixing process thedesign of the infusion catheter plays an eminent role. The in-vitro model was usedto evaluate and validate the technique for several infusion catheter designs, infusionrates, and sensor positions.

Absolute coronary flow rate could be measured reliably over the entire physiologic

General discussion and conclusions 99

range using the continuous infusion thermodilution method. These results werefound under the conditions that a suitable, specially designed, infusion catheter wasused at a high infusion rate of 25 ml/min and positioned in an area with a complexflow pattern. The sensor position should be chosen slightly distal to the mixing zone,which was found to be located 3 to 5 cm downstream of the injection site.

Safety and applicability of the continuous infusion thermodilution technique in anin-vivo model was investigated in chapter 6. Five mongrel dogs were instrumentedwith a perivascular flow probe to obtain true reference flow and with a perivascularballoon occluder to obtain a range of coronary blood flow values. Blood flow valuesdetermined with the continuous infusion thermodilution method were comparedto the true reference flow measured by a perivascular flow probe. Continuousinfusion thermodilution could be used safely during catheterization to obtain maximalcoronary blood flow in dogs. The methodology was found to be feasible, reproducibleand fairly accurate.

Next, the technique was applied in humans which has been described in chapter7. As an indirect measure for agreement, improvement in flow determined bythe thermodilution technique before and after stenting was compared to valuesfor flow improvement based on pressure measurements. In line with the animalstudy, continuous infusion thermodilution could be used safely during catheterization.Moreover, the measurements were reproducible and good agreements were foundwith the indirect validation method. Being able now to determine absolute coronaryblood flow together with distal coronary pressure, it would be possible to calculateabsolute resistance of the coronary artery and of the coronary microcirculation. Animportant issue in this respect is that, by simultaneous measurement of FFR, themeasured value of absolute flow can be related to the virtual normal value as it shouldbe in case of a completely normal coronary artery and microcirculation. The findingsof these studies have been applied already in patients undergoing stem-cell therapyand will be used in heart transplant patients soon.

Chapters 5, 6, and 7 describe the entire process from a clinical desire to selectivelymeasure coronary blood flow towards the validated technique that can now be usedduring cardiac catheterisation. The theoretical background of the thermodilutiontechnique, that is beyond the knowledge and skills of the medical doctor, wastranslated into a comprehensive and clinically usable method. The application ofthe method required an infusion catheter that was developed during the course of theexperiments. Both clinical and technical demands were translated into the catheterdesign. In-vitro tests were designed and executed preceding clinical testing. Forthe animal study performed, thorough and proper knowledge of specific measuringtools belonged to capacities of the medical engineer. Finally, during the clinical studyphase the medical engineer supported, the (sometimes complex) aspects of the studyregarding e.g. the instrumentation, and helped to interpret these study results.

Hemodynamic measurements together with commonly used diagnostic toolsyields a profitable combination as was described in chapter 8, investigating thelarge peripheral arterial system, the domain of the vascular surgeon. In this chapter

100 Chapter 9

invasive pressure measurements within an abdominal aortic aneurysm were carriedout simultaneously with dynamic MRI scans in ten patients, which comprised atechnical challenge. For the first time distensibility and compliance of the wall of theaneurysm have been determined in humans by these simultaneous intra-aneurysmalpressure and volume measurements. The acquisition, analysis, and interpretation ofthe data required the expertise of the medical engineer. MRI-based monitoring of thebio-mechanical parameters determined in this study may become useful for onlinerupture risk assessment and follow-up of aneurysms.

9.2 The Medical Engineer

The research questions that had arisen from clinical practice were translated into thedescribed projects in collaboration with the cardiovascular departments. In additionto the answers to these practical research questions, the project clarifies the role ofthe medical engineer in a non-academic cardiovascular center.

The medical engineer has been involved in initiating and executing the separateresearch projects. For each project a deliberate study design and hemodynamicmeasurement methods were made and data analysis was performed, which primarilyfocussed to answer the practical clinical research question. In the design of theclinical research protocols the possibility of interference with daily practice, ethicalissues, and considerations of accuracy versus applicability of the research methodswere taken into account. All experiments have been executed without any problemregarding the latter issues as a result of the proper planning and communicationwith the involved departments. Moreover, each project performed was a relevantcontribution to clinical knowledge, reflected by the chapters in this thesis which havebeen accepted and published in peer reviewed journals.

For most medical doctors precise understanding and description of the physicalprocesses are subordinate to the practical implications of this research. The medicalengineer negotiates between scientific robustness and practical limitations to a studysetting in clinical practice. The knowledge of the clinical process on the one hand andthe acquaintance with laboratory experiments on the other hand give the medicalengineer an additive value for both clinical and fundamental research. In additionto the ability to function as an intermediary between clinical research and morefundamental science, the same competences are desirable towards (bio-)medicalcompanies in the development of devices and software. Although not specificallyaddressed in this work, the final design of the infusion catheter used for thecontinuous infusion thermodilution technique has been realized in close collaborationwith such a company.

Besides the described work, the medical engineer has been engaged in otherresearch projects at the cardiovascular departments supporting the clinical doctorsin a number of ways. Practical and theoretical support was given among othersfor building and maintaining databases, design of research protocols, and statisticalanalyses. Additionally, the medical engineer has shown the capacity to stimulate and

General discussion and conclusions 101

initiate new research on the cutting edge of life science and engineering as well.

9.3 Conclusion

Besides the conclusions that ensued from the separate projects performed at thedepartments of cardiology, cardiothoracic surgery, and vascular surgery, this workreflects the role of the medical engineer in a non-academic heart center. The medicalengineer should be capable to initiate, stimulate, and execute clinical experimentalresearch, most of which focussed on practical answers. Suitable models, experiments,or protocols can be designed to obtain detailed information of physical processes orto help make a diagnosis. The medical engineer is a valuable addition to a clinicaldepartment.

102 Chapter 9

References

W. Aarnoudse. Invasive assessment of the coronary microcirculation by pressure andtemperature measurements. PhD thesis, Eindhoven University of Technology, 2006.

W. Aarnoudse, W. F. Fearon, G. Manoharan, M. C. F. Geven, F. N. van de Vosse,M. C. M. Rutten, B. De Bruyne, and N. H. J. Pijls. Epicardial stenosis severity doesnot affect minimal microcirculatory resistance. Circulation, 110:2137–2142, 2004a.

W. Aarnoudse, P. van den Berg, F. N. van de Vosse, M. C. F. Geven, M. C. M. Rutten,M. van Turnhout, W. F. Fearon, B. De Bruyne, and N. H. J. Pijls. Myocardialresistance assessed by guidewire-based pressure-temperature measurement: invitro validation. Catheter Cardiovasc Interv, 62:56–63, 2004b.

W. Aarnoudse, M. van’t Veer, N. H. J. Pijls, J. Ter Woorst, S. Vercauteren, P. A. L.Tonino, M. C. F. Geven, M. C. M. Rutten, E. van Hagen, B. de Bruyne, and F. N.van de Vosse. Direct volumetric blood flow measurement in coronary arteries bythermodilution. J. Am. Coll. Cardiol., 50(24):2294–2304, 2007.

F. Alfonso, J. Botas, and eds. Intracoronary Diagnostic Techniques. McGraw-Hill,Madrid, 2003.

F. R. Arko, E. H. Murphy, III C. M. Davis, E. D. Johnson, S. T. Smith, and C. K.Zarins. Dynamic geometry and wall thickness of the aortic neck of abdominalaortic aneurysms with intravascular ultrasonography. J. Vasc. Surg., 46(5):891–896, 2007.

F. Bakhtiary, M. Schiemann, O. Dzemali, T. Wittlinger, M. Doss, H. Ackermann,A. Moritz, and P. Kleine. Stentless bioprostheses improve postoperative coronaryflow more than stented prostheses after valve replacement for aortic stenosis. J.Thorac. Cardiovasc. Surg., 131(4):883–888, 2006.

E. Barbato, W. Aarnoudse, W. R. Aengevaeren, G. Werner, V. Klauss, W. Bojara,I. Herzfeld, K. G. Oldroyd, N. H. J. Pijls, and B. de Bruyne. Validation of coronaryflow reserve measurements by thermodilution in clinical practice. Eur. Heart J, 25(3):219–223, 2004.

J. Bartunek, W. Wijns, G. R. Heyndrickx, and B. de Bruyne. Effects of dobutamineon coronary stenosis physiology and morphology: comparison with intracoronaryadenosine. Circulation, 100(3):243–249, 1999.

J. M. Bland and D. G. Altman. Statistical methods for assessing agreement betweentwo methods of clinical measurement. Lancet, 1(8476):307–310, 1986.

103

104 References

A. Bol, J. A. Melin, J. L. Vanoverschelde, T. Baudhuin, D. Vogelaers, M. DePauw, C. Michel, A. Luxen, D. Labar, M. Cogneau, and . Direct comparisonof [13n]ammonia and [15o]water estimates of perfusion with quantification ofregional myocardial blood flow by microspheres. Circulation, 87(2):512–525,1993.

L. C. Brown and J. T. Powell. Risk factors for aneurysm rupture in patients keptunder ultrasound surveillance. uk small aneurysm trial participants. Ann. Surg.,230:289–296, 1999.

M. Buchi and R. Jenni. Measurement of flow velocity in the coronary circulation:requirements and pitfalls. Semin. Interv. Cardiol., 3(1):45–50, 1998.

P. G. Camici and F. Crea. Coronary microvascular dysfunction. N. Engl. J Med., 356(8):830–840, 2007.

S. G. Carlier, L. C. van Damme, C. P. Blommerde, J. J. Wentzel, G. van Langehove,S. Verheye, M. M. Kockx, M. W. Knaapen, C. Cheng, F. Gijsen, D. J. Duncker,N. Stergiopulos, C. J. Slager, P. W. Serruys, and R. Krams. Augmentation of wallshear stress inhibits neointimal hyperplasia after stent implantation: inhibitionthrough reduction of inflammation? Circulation, 107(21):2741–2746, 2003.

A. J. Carter, W. Wei, L. Gibson, R. Collingwood, F. Tio, J. Dooley, and G. A. Kopia.Segmental vessel wall shear stress and neointimal formation after sirolimus-elutingstent implantation: physiological insights in a porcine coronary model. Cardiovasc.Revasc. Med., 6(2):58–64, 2005.

W. M. Chilian, S. M. Layne, E. C. Klausner, C. L. Eastham, and M. L. Marcus.Redistribution of coronary microvascular resistance produced by dipyridamole. Am.J. Physiol., 256:H383–H390, 1989.

K. P. Conway, J. Byrne, M. Townsend, and I. F. Lane. Prognosis of patients turneddown for conventional abdominal aortic aneurysm repair in the endovascular andsonographic era: Szilagyi revisited? J. Vasc. Surg., 33(4):752–757, 2001.

R. C. Darling, C. R. Messina, D. C. Brewster, and L. W. Ottinger. Autopsy study ofunoperated abdominal aortic aneurysms. the case for early resection. Circulation,56:II161–II164, 1977.

B. de Bruyne, J. Bartunek, S. U. Sys, N. H. J. Pijls, G. R. Heyndrickx, and W. Wijns.Simultaneous coronary pressure and flow velocity measurements in humans.feasibility, reproducibility, and hemodynamic dependence of coronary flow velocityreserve, hyperemic flow versus pressure slope index, and fractional flow reserve.Circulation, 94(8):1842–1849, 1996.

B. de Bruyne, F. Hersbach, N. H. J. Pijls, J. Bartunek, J. W. Bech, G. R. Heyndrickx,K. L. Gould, and W. Wijns. Abnormal epicardial coronary resistance in patientswith diffuse atherosclerosis but ”normal” coronary angiography. Circulation, 104(20):2401–2406, 2001a.

B. de Bruyne, N. H. J. Pijls, J. Bartunek, K. Kulecki, J. W. Bech, H. De Winter, P. VanCrombrugge, G. R. Heyndrickx, and W. Wijns. Fractional flow reserve in patientswith prior myocardial infarction. Circulation, 104(2):157–162, 2001b.

References 105

B. de Bruyne, N. H. J. Pijls, L. Smith, M. Wievegg, and G. R. Heyndrickx. Coronarythermodilution to assess flow reserve: experimental validation. Circulation, 104(17):2003–2006, 2001c.

S. de Putter, B. J. Wolters, M. C. M. Rutten, M. Breeuwer, F. A. Gerritsen, and F. N.van de Vosse. Patient-specific initial wall stress in abdominal aortic aneurysms witha backward incremental method. J. Biomech., 40(5):1081–1090, 2007.

C. di Mario, R. Gil, R. Krams, P. J. de Feyter, and W. Serruys. New invasive techniquesof assessment of the physiological significance of coronary stenoses in humans. Eur.Heart J., 16 Suppl I:104–114, 1995.

R. Erbel, M. Haude, H. W. Hopp, D. Franzen, H. J. Rupprecht, B. Heublein, K. Fischer,P. de Jaegere, P. Serruys, W. Rutsch, and P. Probst. Coronary-artery stentingcompared with balloon angioplasty for restenosis after initial balloon angioplasty.restenosis stent study group. N. Engl. J. Med., 339:1672–1678, 1998.

W. F. Fearon, L. B. Balsam, H. M. Farouque, A. D. Caffarelli, R. C. Robbins, P. J.Fitzgerald, P. G. Yock, and A. C. Yeung. Novel index for invasively assessing thecoronary microcirculation. Circulation, 107:3129–3132, 2003.

W. F. Fearon, W. Aarnoudse, N. H. J. Pijls, B. de Bruyne, L. B. Balsam, D. T. Cooke,R. C. Robbins, P. J. Fitzgerald, A. C. Yeung, and P. G. Yock. Microvascular resistanceis not influenced by epicardial coronary artery stenosis severity: experimentalvalidation. Circulation, 109(19):2269–2272, 2004.

M. F. Fillinger, M. L. Raghavan, S. P. Marra, J. L. Cronenwett, and F. E. Kennedy. Invivo analysis of mechanical wall stress and abdominal aortic aneurysm rupture risk.J. Vasc. Surg., 36(3):589–597, 2002.

M. F. Fillinger, S. P. Marra, M. L. Raghavan, and F. E. Kennedy. Prediction of rupturerisk in abdominal aortic aneurysm during observation: wall stress versus diameter.J. Vasc. Surg., 37(4):724–732, 2003.

G. B. Fiore, M. Grigioni, C. Daniele, G. D’Avenio, V. Barbaro, and R. Fumero.Hydraulic functional characterisation of aortic mechanical heart valve prosthesesthrough lumped-parameter modelling. J. Biomech., 35(10):1427–1432, 2002.

C. Fleming, E. P. Whitlock, T. L. Beil, and F. A. Lederle. Screening for abdominalaortic aneurysm: a best-evidence systematic review for the u.s. preventive servicestask force. Ann. Intern. Med., 142(3):203–211, 2005.

T. Fujiwara. Characteristics of coronary artery blood flow waveforms and themechanism of myocardial ischemia in aortic valve disease. Ann. Thorac. Cardiovasc.Surg., 7(1):1–3, 2001.

T. Fujiwara, A. Nogami, H. Masaki, H. Yamane, S. Matsuoka, H. Yoshida, H. Fukuda,T. Katsumura, and F. Kajiya. Coronary flow velocity waveforms in aortic stenosisand the effects of valve replacement. Ann. Thorac. Surg., 48(4):518–522, 1989.

M. K. Ganten, U. Krautter, H. Tengg-Kobligk, D. Bockler, H. Schumacher, W. Stiller,S. Delorme, H. U. Kauczor, G. W. Kauffmann, and M. Bock. Quantification ofaortic distensibility in abdominal aortic aneurysm using ecg-gated multi-detectorcomputed tomography. Eur. Radiol., 18:966–973, 2008.

106 References

W. Ganz, K. Tamura, H. S. Marcus, R. Donoso, S. Yoshida, and H. J. Swan.Measurement of coronary sinus blood flow by continuous thermodilution in man.Circulation, 44(2):181–195, 1971.

M. C. F. Geven, V. N. Bohte, W. Aarnoudse, P. M. van den Berg, M. C. M. Rutten,N. H. J. Pijls, and F. N. van de Vosse. A physiologically representative in vitromodel of the coronary circulation. Physiol Meas., 25(4):891–904, 2004.

M.C.F. Geven. Methods to assess the coronary circulation by guidewire-mounted sensors.PhD thesis, Eindhoven University of Technology, 2007.

F. J. Gijsen, R. M. Oortman, J. J. Wentzel, J. C. Schuurbiers, K. Tanabe, M. Degertekin,J. M. Ligthart, A. Thury, P. J. de Feyter, P. W. Serruys, and C. J. Slager. Usefulness ofshear stress pattern in predicting neointima distribution in sirolimus-eluting stentsin coronary arteries. Am. J. Cardiol., 92(11):1325–1328, 2003.

H. Glimaker, L. Holmberg, A. Elvin, O. Nybacka, B. Almgren, C. G. Bjorck, andI. Eriksson. Natural history of patients with abdominal aortic aneurysm. Eur. J.Vasc. Surg., 5:125–130, 1991.

K. L. Gould. Identifying and measuring severity of coronary artery stenosis.quantitative coronary arteriography and positron emission tomography. Circulation,78(2):237–245, 1988.

K. L. Gould, R. L. Kirkeeide, and M. Buchi. Coronary flow reserve as a physiologicmeasure of stenosis severity. J. Am. Coll. Cardiol., 15(2):459–474, 1990.

C. E. Hanekamp, J. J. Koolen, N. H. J. Pijls, H. R. Michels, and H. J. Bonnier.Comparison of quantitative coronary angiography, intravascular ultrasound, andcoronary pressure measurement to assess optimum stent deployment. Circulation,99(8):1015–1021, 1999.

G. Hautvast, S. Lobregt, M. Breeuwer, and F. Gerritsen. Automatic contourpropagation in cine cardiac magnetic resonance images. IEEE Trans. Med. Imaging.,25(11):1472–1482, 2006.

D. J. Hildick-Smith and L. M. Shapiro. Coronary flow reserve improves after aorticvalve replacement for aortic stenosis: an adenosine transthoracic echocardiographystudy. J. Am. Coll. Cardiol., 36(6):1889–1896, 2000.

C. Irace, C. Cortese, E. Fiaschi, C. Carallo, E. Farinaro, and A. Gnasso. Wall shear stressis associated with intima-media thickness and carotid atherosclerosis in subjects atlow coronary heart disease risk. Stroke, 35(2):464–468, 2004.

A. Kenny, C. R. Wisbey, and L. M. Shapiro. Profiles of coronary blood flow velocityin patients with aortic stenosis and the effect of valve replacement: a transthoracicechocardiographic study. Br Heart J, 71:57–62, 1994.

M. J. Kern, B. de Bruyne, and N. H. J. Pijls. From research to clinical practice:current role of intracoronary physiologically based decision making in the cardiaccatheterization laboratory. J Am. Coll. Cardiol., 30(3):613–620, 1997.

P. Kleine, M. J. Hasenkam, H. Nygaard, M. Perthel, D. Wesemeyer, and J. Laas. Tiltingdisc versus bileaflet aortic valve substitutes: intraoperative and postoperativehemodynamic performance in humans. J. Heart Valve Dis., 9(2):308–311, 2000.

References 107

P. Kleine, U. Abdel-Rahman, A. A. Klesius, M. Scherer, A. Simon, and A. Moritz.Comparison of hemodynamic performance of medtronic hall 21 mm versus st. judemedical 23 mm prostheses in pigs. J. Heart Valve Dis., 11(6):857–863, 2002a.

P. Kleine, A. A. Klesius, M. Scherer, U. Abdel-Rahman, and A. Moritz. Initial in vivoresults of the new medtronic advantage(tm) bileaflet valve in aortic position andcomparison to the sjm. Cardiovasc. Surg., 10(5):494–499, 2002b.

P. Kleine, M. Scherer, U. Abdel-Rahman, A. A. Klesius, H. Ackermann, and A. Moritz.Effect of mechanical aortic valve orientation on coronary artery flow: comparisonof tilting disc versus bileaflet prostheses in pigs. J. Thorac. Cardiovasc. Surg., 124(5):925–932, 2002c.

D.N. Ku. Blood flow in arteries. Annu. Rev. Fluid. Mech., 29:399–434, 1997.A. Lalande, van K. Khau, N. Salve, D. Ben Salem, L. Legrand, P. M. Walker, J. E. Wolf,

and F. Brunotte. Automatic determination of aortic compliance with cine-magneticresonance imaging: an application of fuzzy logic theory. Invest Radiol., 37(12):685–691, 2002.

R. M. Lang, B. P. Cholley, C. Korcarz, R. H. Marcus, and S. G. Shroff. Measurement ofregional elastic properties of the human aorta. a new application of transesophagealechocardiography with automated border detection and calibrated subclavian pulsetracings. Circulation, 90(4):1875–1882, 1994.

F. A. Lederle, S. E. Wilson, G. R. Johnson, D. B. Reinke, F. N. Littooy, C. W. Acher,D. J. Ballard, L. M. Messina, I. L. Gordon, E. P. Chute, W. C. Krupski, S. J. Busuttil,G. W. Barone, S. Sparks, L. M. Graham, J. H. Rapp, M. S. Makaroun, G. L. Moneta,R. A. Cambria, R. G. Makhoul, D. Eton, H. J. Ansel, J. A. Freischlag, and D. Bandyk.Immediate repair compared with surveillance of small abdominal aortic aneurysms.N. Engl. J. Med., 346(19):1437–1444, 2002.

A. Long, L. Rouet, A. Bissery, P. Rossignol, D. Mouradian, and M. Sapoval. Complianceof abdominal aortic aneurysms evaluated by tissue doppler imaging: correlationwith aneurysm size. J. Vasc. Surg., 42(1):18–26, 2005.

R. Lopez-Palop, E. Pinar, I. Lozano, D. Saura, F. Pico, and M. Valdes. Utility ofthe fractional flow reserve in the evaluation of angiographically moderate in-stentrestenosis. Eur. Heart J., 25(22):2040–2047, 2004.

E. S. Di Martino, G. Guadagni, A. Fumero, G. Ballerini, R. Spirito, P. Biglioli, andA. Redaelli. Fluid-structure interaction within realistic three-dimensional models ofthe aneurysmatic aorta as a guidance to assess the risk of rupture of the aneurysm.Med. Eng Phys., 23:647–655, 2001.

E. S. Di Martino, A. Bohra, J. P. Vande Geest, N. Gupta, M. S. Makaroun, and D. A.Vorp. Biomechanical properties of ruptured versus electively repaired abdominalaortic aneurysm wall tissue. J. Vasc. Surg., 43(3):570–576, 2006.

D. G. Mathey, K. Chatterjee, J. V. Tyberg, J. Lekven, B. Brundage, and W. W. Parmley.Coronary sinus reflux. a source of error in the measurement of thermodilutioncoronary sinus flow. Circulation, 57(4):778–786, 1978.

A. Matrai, R.B. Whittington, and E. Ernst. A simple method of estimating whole bloodviscosity at standardized hematocrit. Clin. Hemorheol., 7:261–265, 1987.

108 References

P. Meier and K. L. Zierler. On the theory of the indicator-dilution method formeasurement of blood flow and volume. J. Appl. Physiol., 6:731–744, 1954.

M. C. Morice, P. W. Serruys, J. E. Sousa, J. Fajadet, Hayashi E. Ban, M. Perin,A. Colombo, G. Schuler, P. Barragan, G. Guagliumi, F. Molnar, and R. Falotico.A randomized comparison of a sirolimus-eluting stent with a standard stent forcoronary revascularization. N. Engl. J. Med., 346(23):1773–1780, 2002.

J. W. Moses, M. B. Leon, J. J. Popma, P. J. Fitzgerald, D. R. Holmes, C. O’Shaughnessy,R. P. Caputo, D. J. Kereiakes, D. O. Williams, P. S. Teirstein, J. L. Jaeger, and R. E.Kuntz. Sirolimus-eluting stents versus standard stents in patients with stenosis in anative coronary artery. N. Engl. J. Med., 349(14):1315–1323, 2003.

A. Nemes, T. Forster, Z. Kovacs, A. Thury, I. Ungi, and M. Csanady. The effect of aorticvalve replacement on coronary flow reserve in patients with a normal coronaryangiogram. Herz, 27(8):780–784, 2002.

S. C. Nicholls, J. B. Gardner, M. H. Meissner, and H. K. Johansen. Rupture in smallabdominal aortic aneurysms. J. Vasc. Surg., 28(5):884–888, 1998.

M. Peltier, F. Trojette, M. E. Sarano, F. Grigioni, M. A. Slama, and C. M. Tribouilloy.Relation between cardiovascular risk factors and nonrheumatic severe calcific aorticstenosis among patients with a three-cuspid aortic valve. Am. J. Cardiol., 91:97–99,2003.

L. H. Peterson, R. E. Jensen, and J. Parnell. Mechanical properties of arteries in vivo.Circ. Res., 8:622–639, 1960.

P. N. Petropoulakis, M. K. Kyriakidis, C. A. Tentolouris, C. V. Kourouclis, and P. K.Toutouzas. Changes in phasic coronary blood flow velocity profile in relation tochanges in hemodynamic parameters during stress in patients with aortic valvestenosis. Circulation, 92(6):1437–1447, 1995.

N. H. J. Pijls. Optimum guidance of complex pci by coronary pressure measurement.Heart, 90(9):1085–1093, 2004.

N. H. J. Pijls and B. DeBruyne. Coronary Pressure. Kluwer Academic Publisher,Dordrecht, 2000.

N. H. J. Pijls, G. J. Uijen, A. Hoevelaken, T. Arts, W. R. Aengevaeren, H. S. Bos,J. H. Fast, K. L. van Leeuwen, and Werf T. van der. Mean transit time for theassessment of myocardial perfusion by videodensitometry. Circulation, 81(4):1331–1340, 1990.

N. H. J. Pijls, J. A. van Son, R. L. Kirkeeide, B. de Bruyne, and K. L. Gould.Experimental basis of determining maximum coronary, myocardial, and collateralblood flow by pressure measurements for assessing functional stenosis severitybefore and after percutaneous transluminal coronary angioplasty. Circulation, 87(4):1354–1367, 1993.

N. H. J. Pijls, G. J. Bech, M. I. el Gamal, H. J. Bonnier, B. de Bruyne, B. van Gelder,H. R. Michels, and J. J. Koolen. Quantification of recruitable coronary collateralblood flow in conscious humans and its potential to predict future ischemic events.J Am. Coll. Cardiol., 25(7):1522–1528, 1995a.

References 109

N. H. J. Pijls, B. van Gelder, P. van der Voort, K. Peels, F. A. L. E. Bracke, H. J. Bonnier,and M. I. el Gamal. Fractional flow reserve. a useful index to evaluate the influenceof an epicardial coronary stenosis on myocardial blood flow. Circulation, 92(11):3183–3193, 1995b.

N. H. J. Pijls, B. de Bruyne, K. Peels, P. H. van der Voort, H. J. Bonnier, J. K. J.Bartunek, and J. J. Koolen. Measurement of fractional flow reserve to assess thefunctional severity of coronary-artery stenoses. N. Engl. J Med., 334(26):1703–1708, 1996.

N. H. J. Pijls, B. de Bruyne, L. Smith, W. Aarnoudse, E. Barbato, J. Bartunek, G. J.Bech, and V de van. Coronary thermodilution to assess flow reserve: validation inhumans. Circulation, 105(21):2482–2486, 2002a.

N. H. J. Pijls, V. Klauss, U. Siebert, E. Powers, K. Takazawa, W. F. Fearon, J. Escaned,Y. Tsurumi, T. Akasaka, H. Samady, and B. de Bruyne. Coronary pressuremeasurement after stenting predicts adverse events at follow-up: a multicenterregistry. Circulation, 105(25):2950–2954, 2002b.

N. H. J. Pijls, P. van Schaardenburgh, G. Manoharan, E. Boersma, J. W. Bech, M. van’tVeer, F. Bar, J. Hoorntje, J. Koolen, W. Wijns, and B. de Bruyne. Percutaneouscoronary intervention of functionally nonsignificant stenosis: 5-year follow-up ofthe defer study. J. Am. Coll. Cardiol., 49(21):2105–2111, 2007.

J. T. Powell and L. C. Brown. The natural history of abdominal aortic aneurysms andtheir risk of rupture. Acta Chir Belg., 101(1):11–16, 2001.

J. H. Reiber, P. W. Serruys, C. J. Kooijman, W. Wijns, C. J. Slager, J. J. Gerbrands, J. C.Schuurbiers, A. den Boer, and P. G. Hugenholtz. Assessment of short-, medium-, andlong-term variations in arterial dimensions from computer-assisted quantitation ofcoronary cineangiograms. Circulation, 71(2):280–288, 1985.

H. A. Richter and C. Mittermayer. Volume elasticity, modulus of elasticity andcompliance of normal and arteriosclerotic human aorta. Biorheology, 21(5):723–734, 1984.

J. D. Rossen and M. D. Winniford. Effect of increases in heart rate and arterial pressureon coronary flow reserve in humans. J. Am. Coll. Cardiol., 21(2):343–348, 1993.

W. Rutishauser, H. Simon, J. P. Stucky, N. Schad, G. Noseda, and J. Wellauer.Evaluation of roentgen cinedensitometry for flow measurement in models and inthe intact circulation. Circulation, 36(6):951–963, 1967.

M. Sanmartin, J. Goicolea, C. Garcia, J. Garcia, A. Crespo, J. Rodriguez, and J. M.Goicolea. [influence of shear stress on in-stent restenosis: in vivo study using 3dreconstruction and computational fluid dynamics]. Rev. Esp. Cardiol., 59(1):20–27,2006.

J. A. E. Spaan. Coronary Blood Flow. Mechanics, Distribution, and Control. KluwerAcademic Publishers, Dordrecht, 1991.

P. H. Stone, A. U. Coskun, S. Kinlay, M. E. Clark, M. Sonka, A. Wahle, O. J. Ilegbusi,Y. Yeghiazarians, J. J. Popma, J. Orav, R. E. Kuntz, and C. L. Feldman. Effect ofendothelial shear stress on the progression of coronary artery disease, vascularremodeling, and in-stent restenosis in humans: in vivo 6-month follow-up study.Circulation, 108(4):438–444, 2003.

110 References

M. Takeuchi and Y. Nakashima. Effect of aortic valve replacement on coronary flowvelocity during metabolic stress in a patient with aortic stenosis. Cathet. Cardiovasc.Diagn., 40(3):287–290, 1997.

The-UK-Small-Aneurysm-Trial-Participants. Long-term outcomes of immediate repaircompared with surveillance of small abdominal aortic aneurysms. N. Engl. J Med.,346(19):1445–1452, 2002.

The-UK-Small-Aneurysm-Trial-Participants. Mortality results for randomised con-trolled trial of early elective surgery or ultrasonographic surveillance for smallabdominal aortic aneurysms. Lancet, 352(9141):1649–1655, 1998.

L. M. van Bortel, E. J. Balkestein, van der Heijden-Spek JJ, F. H. Vanmolkot, J. A.Staessen, J. A. Kragten, J. W. Vredeveld, M. E. Safar, H. A. Struijker Boudier, andA. P. Hoeks. Non-invasive assessment of local arterial pulse pressure: Comparisonof applanation tonometry and echo-tracking. J. Hypertens., 19:1037–1044, 2001.

M. van’t Veer, N. H. J. Pijls, W. Aarnoudse, J. J. Koolen, and F. N. van deVosse. Evaluation of the haemodynamic characteristics of drug-eluting stents atimplantation and at follow-up. Eur. Heart J, 27(15):1811–1817, 2006.

D. A. Vorp, W. A. Mandarino, M. W. Webster, and III J. Gorcsan. Potential influenceof intraluminal thrombus on abdominal aortic aneurysm as assessed by a new non-invasive method. Cardiovasc. Surg., 4:732–739, 1996.

D. A. Vorp, M. L. Raghavan, and M. W. Webster. Mechanical wall stress in abdominalaortic aneurysm: influence of diameter and asymmetry. J. Vasc. Surg., 27(4):632–639, 1998.

A. B. Weisse and T. J. Regan. A comparison of thermodilution coronary sinus bloodflows and krypton myocardial blood flows in the intact dog. Cardiovasc. Res., 8(4):526–533, 1974.

R. Wentz, A. Manduca, J. G. Fletcher, H. Siddiki, R. C. Shields, T. Vrtiska, G. Spencer,A. N. Primak, J. Zhang, T. Nielson, C. McCollough, and L. Yu. Automaticsegmentation and co-registration of gated ct angiography datasets: measuringabdominal aortic pulsatility. In Proceedings of the SPIE, volume 6511, pages65111I1–I9, 2007.

J. J. Wentzel, R. Krams, J. C. Schuurbiers, J. A. Oomen, J. Kloet, W. J. Der Giessen,P. W. Serruys, and C. J. Slager. Relationship between neointimal thickness andshear stress after wallstent implantation in human coronary arteries. Circulation,103(13):1740–1745, 2001.

C. W. White. Clinical applications of doppler coronary flow reserve measurements.Am. J. Cardiol., 71(14):10D–16D, 1993.

K. Wilson, H. MacCallum, I. B. Wilkinson, P. R. Hoskins, A. J. Lee, and A. W.Bradbury. Comparison of brachial artery pressure and derived central pressure inthe measurement of abdominal aortic aneurysm distensibility. Eur. J. Vasc. Endovasc.Surg., 22(4):355–360, 2001.

K. A. Wilson, A. J. Lee, A. J. Lee, P. R. Hoskins, F. G. Fowkes, C. V. Ruckley, andA. W. Bradbury. The relationship between aortic wall distensibility and rupture ofinfrarenal abdominal aortic aneurysm. J. Vasc. Surg., 37(1):112–117, 2003.

References 111

R. F. Wilson, D. E. Laughlin, P. H. Ackell, W. M. Chilian, M. D. Holida, C. J. Hartley,M. L. Armstrong, M. L. Marcus, and C. W. White. Transluminal, subselectivemeasurement of coronary artery blood flow velocity and vasodilator reserve in man.Circulation, 72(1):82–92, 1985.

K.L. Ziehler. In: Circulation times and the theory of indicator dilution methods fordetermining blood flow and volume. American Physiology Society, Washington, DC,1962.

D. P. Zipes, P. Libby, R. O. Bonow, and E. Braunwald. Baunwald’s Heart Disease. Atextbook of cardiovascular medicine. Elsevier Saunders, Philidelphia, Pennsylvania,7th edition, 2005.

112 References



Summary

The main cause of the degeneration of the cardiovascular system is atherosclerosisand the diagnosis of pathology related to atherosclerosis is often based on anatomicmeasurements alone. Functional severity cannot be determined based these anatomicdata and requires hemodynamic measurements. Hemodynamic measurementsthroughout the cardiovascular system are the main focus of this work. A generalintroduction discussing the aims and outline of this thesis is presented in chapter 1.The consequences of degeneration resulting from atherosclerosis on particular aspectsof the cardiovascular system, hereby introducing the separate chapters of this work,are discussed shortly in chapter 2.

Percutaneous coronary intervention is often performed to treat coronary arterydisease. To reopen the coronary artery a variety of stents might be used. In chapter3, two types of stents, a bare metal stent and a drug-eluting stent, were comparedwith respect to anatomic and physiologic measurements. The stents were implantedin pairs within the same patient to exclude differences as a result of the biologicdiversity between patients. No differences were obtained just before and just after thestent-implantation. At six-month follow-up, however, the drug-eluting stent showedsuperior results compared to its bare metal counterpart, not only with respect toanatomic but also with respect to physiologic characteristics.

It was found that, in contrast to the bare metal stent, the sirolimus stentmaintained a normal wall shear stress value within the stented segment. This globalwall shear stress estimation is a better indicator for the local hemodynamics withinthe stent than the anatomy-derived parameters alone because it accounts for theaverage cross-sectional velocity as well. Pressure derived fractional flow reservewas significantly higher and the hyperemic trans-stent gradient significantly lower

113

114 Summary

for the drug eluting stent. Theoretically, the latter finding improves the possibilitiesof interventional treatment of patients with multiple non-adjacent lesions within oneartery. However, despite the superiority based on anatomic and hemodynamic results,the increased risk of (sub-) acute in-stent thrombosis asks for a deliberate choice forusing multiple drug-eluting stents. Irrespective of the consequences of the results, itcan be concluded that the drug-eluting sirolimus stent is superior to its bare metalcounterpart with respect to hemodynamic characteristics at follow-up.

Physiologic measurements have been used in chapter 4 as well to determine theinfluence of the orientation of a bi-leaflet valve prosthesis on coronary perfusion. Itwas hypothesized that coronary blood flow was altered by the ejection velocity ofthe blood through the mechanical valve prosthesis. This influence was assumed tobe dependent on the orientation of the valve prosthesis. Therefore, three groups ofpatients were included in the study and compared with respect to coronary perfusionpressure. Patients receiving a mechanical bi-leaflet prosthesis were randomized toeither the orientation with the hinge mechanism parallel to a line drawn betweenthe coronary ostia, or the orientation with the hinge mechanism perpendicular to aline drawn between the coronary ostia. Patients receiving bio-prosthesis comprisedthe control group. Six months after aortic valve replacement a catheterisation wasperformed and pressures were measured at several spots in the aorta and underseveral circumstances.

From the results of this study it can be concluded that the influence of a bi-leaflet prosthesis on coronary perfusion pressure was negligible and not dependent onvalve orientation. Moreover, coronary perfusion pressures did not differ significantlybetween mechanical and biological prostheses.

Chapters 3 and 4 have focussed on coronary physiology of the epicardialarteries. To interrogate the coronary microvasculature, the ability to measure absolutecoronary blood flow would be desirable. In chapters 5, 6, and 7 a technique based oncontinuous infusion thermodilution has been developed and validated. Theoretically,absolute coronary blood flow through a selective coronary artery could be calculatedfrom the temperature of the blood, the saline, and the mixture downstream of theinjection site, combined with a known infusion rate. Application of the technique toobtain reliable values for coronary blood flow requires the blood and the saline tobe completely mixed downstream of the site of injection. In the mixing process thedesign of the infusion catheter plays an eminent role. An in-vitro model is describedin chapter 5 to validate the technique for several infusion catheter designs, infusionrates, and sensor positions.

Safety and applicability of the continuous infusion thermodilution techniquewas tested in animals (chapter 6). Five mongrel dogs were instrumented with aperivascular flow probe to obtain true reference flow. During the experiments aspecially designed infusion catheter, as described in chapter 5, was used. Bloodflow values determined with the continuous infusion thermodilution method werecompared to the true reference flow measured by the perivascular flow probe. Next,the technique was applied in humans which is described in chapter 7.

Summary 115

Coronary blood flow could be measured safely in selective coronary arteries duringcardiac catheterisation. Continuous infusion thermodilution is fairly reliable if a suit-able, specially designed, infusion catheter was used. Being able to determine absolutecoronary blood flow together with distal coronary pressure, it is possible to calculateabsolute resistance of the coronary artery and of the coronary microcirculation.

Hemodynamic measurements together with commonly used diagnostic toolsyields a profitable combination as was described in chapter 8. In this chapter invasivepressure measurements within an abdominal aortic aneurysm were carried outsimultaneously with dynamic MRI scans in ten patients. For the first time distensibilityand compliance of the wall of the aneurysm were determined in humans by thesesimultaneous intra-aneurysmal pressure and volume measurements. Furthermore,brachial cuff measurements were different compared to the invasive measurementswithin the aneurysm. Therefore, non-invasive measurements can only be used toobtain a relative parameter rather than an absolute values for distensibility. MRI-based monitoring of this bio-mechanical parameter may become useful for onlinerupture risk assessment and follow-up of aneurysms in combination with brachialpressure measurements.

Finally, in chapter 9 a general discussion is presented and conclusions are drawnon the role of the medical engineer in a large non-academic heart center, which wasthe overall goal of this thesis. The several research projects in this thesis have shownthe additive value of the medical engineer in a cardiovascular department of a non-academic heart center.

116 Summary

Samenvatting

Atherosclerose of aderverkalking is de meest voorkomende oorzaak van de degener-atie van de bloedvaten. Als het vaatstelsel aangedaan is, wordt de diagnose vaakgesteld door enkel naar anatomische kenmerken te kijken. Echter, in hoeverrede functie van het betreffende deel van het vaatstelsel is aangedaan, kan nietworden afgeleid uit deze anatomische gegevens. Om dit te bepalen zijn lokalemetingen van bloeddruk en bloedstroom nodig, ofwel hemodynamische metingen.Het uitvoeren van deze metingen op verschillende plaatsen in het vaatstelsel engebruik van deze metingen voor een betere diagnostiek vormt het belangrijkste deelvan dit proefschrift. Tevens wordt in dit proefschrift duidelijk wat de rol is vande medisch ingenieur in een groot niet-academisch hartcentrum. In hoofdstuk 1wordt een algemene inleiding gegeven met het doel en indeling van de verschillendehoofdstukken van dit proefschrift. In hoofdstuk 2 wordt vervolgens, voor een aantalspecifieke onderdelen van het vaatstelsel, besproken wat de consequenties van hetdegenereren van het vaatstelsel zijn voor een patient. Bovendien wordt besprokenhoe hemodynamische metingen een toegevoegde waarde kunnen hebben bij dehuidige diagnostiek.

Dotteren, ofwel een percutane coronaire interventie, is een veel uitgevoerdebehandeling om vernauwingen in de kransslagaders op te heffen. In de meestegevallen wordt bij dotteren een stent in het bloedvat achtergelaten. Dit soortbuisjes zijn er in veel varianten. In hoofdstuk 3 zijn twee varianten van stentsmet elkaar vergeleken: een drug-eluting stent en een bare-metal stent. De stentswerden paarsgewijs geplaatst in dezelfde patient om zodoende de omstandighedenzo vergelijkbaar mogelijk te maken. Net na implantatie werd er geen verschilgevonden tussen de metingen voor beide stents. Echter, bij het her-onderzoek, zesmaanden na implantatie, waren de resultaten voor de drug-eluting stent beter danvoor de bare-metal stent met betrekking tot zowel de anatomische als de fysiologischekarakteristieken.

Bovendien werd er, in tegenstelling tot de bare-metal stent, een homogeneen normale wandschuifspanning gevonden in de drug-eluting stent. Deze globaleschatting voor de wandschuifspanning, zoals gebruikt in deze studie, geeft een betereindruk van de lokale hemodynamica dan de anatomische metingen. Bij de berekeningvan de wandschuifspanning wordt naast de anatomische parameters ook bloedstroomsnelheid meegenomen. Ook werd gevonden dat de fractionele flow reserve en de

117

118 Samenvatting

trans-stent drukgradient bij maximale bloedstroom beter waren voor de drug-elutingstent dan voor de bare-metal stent. Uit dit onderzoek kan geconcludeerd wordendat de drug-eluting stent superieur is vergeleken met de bare-metal stent. Ondanksde anatomisch en fysiologisch betere resultaten voor de drug-eluting stent moeteen weloverwogen keuze gemaakt worden om meerdere van dit soort stents in eenkransslagader te plaatsen met het oog op het risico van (sub-) acute stent trombose.

In hoofdstuk 4 zijn soortgelijke hemodynamische metingen uitgevoerd om teonderzoeken of de orientatie van een mechanische hartklepprothese invloed heeftop de doorstroming van de kransslagaders. Er is in het verleden verondersteld datde bloedstroom door de hartklepprothese de perfusie van de kransslagaders, die vlakboven de aortaklep ontspringen, kon beınvloeden (venturi-effect). Bovendien werdverondersteld dat de invloed afhankelijk zou zijn van de orientatie van de prothese.Om dit te onderzoeken werden drie groepen patienten gevraagd mee te werken aanhet onderzoek. Bij de eerste groep werd de klepprothese dusdanig geplaatst dathet scharnier-mechanisme parallel liep aan een denkbeeldige lijn getrokken tussende twee coronaire ostia. Bij de tweede groep werd de prothese een kwartslaggedraaid ten opzichte van de eerst genoemde orientatie. De derde groep patientenfungeerde als een controle groep. Deze patenten kregen een biologische prothese.Zes maanden na de aortaklep vervanging werd een hartcatheterisatie uitgevoerd enwerden drukmetingen gedaan op verschillende plaatsen in de aorta in rust en tijdensinspanning.

Uit de resultaten kan geconcludeerd worden dat de perfusiedruk van de krans-slagaders nauwelijks werd beınvloed door de aanwezigheid van een mechanischeklepprothese. Ook werd er geen verschil tussen de orientaties gevonden of tussende mechanische en biologische protheses.

Hoofdstukken 3 en 4 richten zich op fysiologische metingen in de epicardialekransslagaders. Om de microvasculatuur van het hart te onderzoeken zou hetwenselijk zijn om een techniek te hebben die de bloedstroom door de kransslagaderskan meten. Dit was nog niet mogelijk tijdens een hartcatheterisatie. In dehoofdstukken 5, 6 en 7 werd een techniek ontwikkeld en gevalideerd die gebaseerdis op temperatuurmetingen tijdens continue infusie van een zoutoplossing in debloedstroom (thermodilutie). In theorie kan de bloedstroom door een kransslagaderberekend worden door de temperatuur van het bloed en van de zoutoplossing temeten samen met de temperatuur van het mengsel van deze twee vloeistoffen. Boven-dien moet de infusiesnelheid van de zoutoplossing bekend zijn. Om nauwkeurigewaarden te vinden voor de bloedstroom is het een vereiste voor deze techniek omvolledige menging te hebben van bloed en de zoutoplossing. Het ontwerp van eeninfusiecatheter speelt een belangrijke rol in het mengproces. In hoofdstuk 5 is eenin-vitro model gebruikt om naast verschillende infusiecatheters, die we ontwikkeldhebben, ook verschillende meetposities en infusiesnelheden te onderzoeken.

In hoofdstuk 6 zijn dierexperimenten uitgevoerd om de veiligheid en toepas-baarheid alsmede de nauwkeurigheid van de techniek in-vivo te onderzoeken.Vijf honden werden geınstrumenteerd met een peri-vasculaire flowmeter om de

Samenvatting 119

werkelijke bloedstroom te meten en een peri-vasculair ballonnetje om de bloedstroomte kunnen beınvloeden. De bloedstroom waarden gebaseerd op de temperatuurmetin-gen werden vergeleken met de werkelijk gemeten bloedstroom met de peri-vasculaireflowmeter. Na deze experimenten werd de techniek toegepast bij patienten die eenhartcatheterisatie ondergingen. Dit is beschreven in hoofdstuk 7.

Uit de resultaten van deze hoofdstukken blijkt dat de bloedstroom door eenkransslagader veilig bepaald kan worden tijdens een hartcatheterisatie. Bovendienis de thermodilutie-techniek redelijk nauwkeurig mits de speciaal ontworpen infusiecatheter gebruikt wordt waarmee goede menging van bloed en zoutoplossing wordtbereikt. Het bepalen van absolute waarden van bloedstroom in een kransslagadermaakt het mogelijk om ook absolute waarden van bloedstroom in de kleinere bloed-vaten te bepalen, de microcirculatie, alsmede de relatieve en absolute microvasculaireweerstand te berekenen.

De combinatie van hemodynamische metingen en de gangbare diagnostischemiddelen blijken een nuttige combinatie te zijn zoals beschreven is in hoofdstuk8. In dit hoofdstuk is beschreven hoe in tien patienten invasieve druk metingenin een aneurysma van de aorta abdominalis worden uitgevoerd simultaan meteen dynamische MRI scan. Door deze metingen is het voor het eerst mogelijkom distensibiliteit en compliantie van de wand van een aneurysma te bepalen opbasis van de gelijktijdige druk en volumegegevens. Bovendien werd tijdens ditonderzoek een niet-invasive bloeddrukmeting aan de bovenarm vergeleken met deinvasief gemeten bloeddruk. De waarden deze twee methoden waren verschillend.De niet-invasieve bloeddrukmetingen kunnen om deze reden niet gebruikt wordenom absolute waarden voor bio-mechanische parameters te bepalen. Echter, dedrukmetingen aan de bovenarm kunnen in combinatie met de dynamische MRIscan wellicht gebruikt worden om veranderingen in bio-mechanische parameters tedetecteren.

Hoofdstuk 9 van dit proefschrift vormt de algemene discussie en conclusies.Hierin worden ook conclusies getrokken over de rol van de medisch ingenieur ineen groot niet-academisch hartcentrum.

120 Samenvatting

Dankwoord

In de afgelopen vier jaar heb ik met plezier uitgekeken naar het moment dat ik hetdankwoord mocht gaan schrijven om twee redenen. Ten eerste om het feit dat hetdan echt zover is dat er een mijlpaal bereikt is; het is af! Ten tweede omdat ditmeest gelezen hoofdstuk bij uitstek een mogelijkheid is om een aantal mensen nogeens persoonlijk te bedanken die hebben bijgedragen aan het tot stand komen van ditproefschrift of anderszins belangrijk zijn geweest.

In de eerste plaats natuurlijk mijn promotoren Nico Pijls en Frans van de Vosse.Beste Nico, ik wil je heel erg bedanken voor het feit dat je mij de mogelijkheidgeboden hebt om als medisch ingenieur een afstudeerproject bij je te doen watvervolgd werd in een promotieonderzoek. Je hebt me wegwijs gemaakt op deverschillende afdelingen en me enorm veel geleerd op het gebied van het doenvan onderzoek. Beste Frans, bedankt voor je altijd nuchtere en kritische kijk op devraagstukken die ik je tijdens mijn promotie voorlegde. Je wist de hobbeltjes op hetonderzoekspad altijd dusdanig te relativeren dat ik me weleens afvroeg waarom ik zein eerste instantie als probleem zag. Ik kijk uit naar de samenwerking met jullie in dekomende jaren.

Maartje Geven en Wilbert Aarnoudse. Beste Maartje en Wilbert, bedankt datik me mocht mengen in jullie onderzoeken. We hebben veel lol gehad tijdens hetvoorbereiden en uitvoeren van de continue infusie experimenten in het (cath)lab.Niet minder belangrijk en zeker zo gezellig waren de afspraken buiten werktijd.

Pim Tonino. Beste Pim, van harte bedankt voor de hand- en span diensten dieje altijd belangeloos deed, voor alle zinvolle en zinloze - met gekrabbel op het whiteboard ondersteunde - discussies, en voor de gezellige tijd tijdens werkgerelateerdebezoekjes aan bijvoorbeeld een golfcourse vlakbij Glasgow.

Mijn copromotor Marcel Rutten. Beste Marcel, ondanks het feit dat je pas ineen laat stadium mijn copromotor werd, ben je betrokken geweest bij alle in-vitroexperimenten die (zijdelings) te maken hadden met het promotieonderzoek. Dankvoor je hulp en ondersteuning hiermee en dank voor het delen van je kennis over deItaliaanse en de Brabantse taal.

Maarten Merkx. Beste Maarten, het werk van je afstuderen is niet directgerelateerd geweest aan het onderzoek in dit proefschrift, maar we hebben er welmooi een artikel van gemaakt; top!

Zonder de gastvrijheid en hulp van een groot aantal mensen van verschillendeafdelingen was ik waarschijnlijk nu nog steeds patienten aan het includeren en

121

122 Dankwoord

gegevens aan het verzamelen:Antoinette, Ingrid, Harold, Hannie, Linda, Anouk, Krisztina, Kathinka, Anne,

Lisette, Monique, Annie en Laura van de R&D en het secretariaat van de Cardiologie:enorm bedankt voor het realiseren van de randvoorwaarden die het doen vanonderzoek uberhaupt mogelijk maken.

Ik ben zeer veel dank verschuldigd aan een ieder op en rond de HCK: Eduard,Henk, Jan, Ruud, Peter, Boudewijn, Arjen, Alex, Marie Jose, Tessa, Kees, Berry,Gert, Bonnie, Roy, Hans, Hans, Claudia, Hanneke, Rudi, Petra, Ingrid, Nicole, Petry.Bedankt voor de geweldig goede sfeer en dat ik te allen tijde bij jullie langs kankomen.

De steun en ondersteuning van een groot aantal mensen van de functie afdelingen van de afdeling radiologie is werkelijk onmisbaar geweest. Marleen en Cindy wilik speciaal bedanken voor hun flexibiliteit en bereidheid om met propvolle schema’sdusdanig te schuiven zodat er een uitgebreide -en soms zeer frustrerende- MRI scanvoor een studie ingepast kon worden.

Onmisbaar waren ook de assistenten van de cardiologie. Jullie hebben mij vanafhet eerste moment het gevoel gegeven dat ik ”part of the family” was. Dank dat julliemij op sleeptouw hebben genomen.

Het werken in het Catharina ziekenhuis aan de projecten zoals beschreven in ditproefschrift waren nooit van de grond gekomen zonder de steun en ideeen van demaatschappen cardiologie, cardiochirurgie en vaatchirurgie. Jacques Koolen, JanMelle van Dantzig, Rolf Michels, Kathinka Peels, Hans Bonnier, Frank Bracke, AlbertMeijer, Kees-Joost Botman, Hans Post, Guus Brueren, Lucas Dekker, Pepijn van deVoort, Bart van Straten, Joost ter Woorst en Jaab Buth: bedankt!

A major part of this thesis comprises measurements of the coronary physiology.The indefatigable enthusiasm of RADI and coworkers as well as the knowledge anddevotion in performing research makes me feel privileged to be part this researchteam. Besides a lot of people I want to say thanks to Huub, Bert, Emiel, Lars, Leif,Johan, Mona, Eva, and Marc. Thank you for a lot of things!

Behalve al deze hulp en ondersteuning op het werk zijn er daarnaast gelukkigook nog mensen die over hele andere dingen kunnen praten dan werk alleen. Heelontspannend zijn de weekendjes en potjes squash met Alina, Inge, Homme-Auke enLambert: Bedankt! Lambert, ook nog een boel bedankjes voor de beschikbaarheidvan jullie ”hotel” en het feit dat je mijn paranimf wilde zijn.

De dank aan de mannen van Heren 1 van ODI Middelbeers is niet in woorden teomschrijven. Jullie hebben een heel speciaal plekje in mijn hart.

Familie is alles. Mariska, Michael, ik ben enorm blij jullie als zus en broeder tehebben. Bedankt voor de hulp met de voorkant (Maris) en dat je paranimf wildezijn (Mike). Onvoorwaardelijke steun, liefde en hulp in alles; pa en ma, bedankt, ditboekje is voor jullie. Lieve Natasja. Met jou kan ik de wereld aan; laten we ’m danmaar samen gaan verkennen . . .

MarcelMiddelbeers, oktober 2008

Curriculum Vitae

Marcel van ’t Veer was born in Soest on June 2nd 1980. He attended secondaryeducation at the Griftland College in Soest, The Netherlands and graduated in 1998.In the same year he started studying Biomedical Engineering at the EindhovenUniversity of Technology and obtained his Master’s degree in Medical Engineeringin 2004. He was one of the first five Medical Engineers in Europe. During his Master’straining he spent three months at Stanford University, Palo Alto, California, USA inthe laboratory of Dr. Charles A. Taylor where he studied flow patterns in and arounda stented coarctation.

For his Master’s thesis he performed research at the department of cardiologyof the Catharina Hospital in Eindhoven. After he obtained his Master’s degree theresearch was continued at the department of cardiology with topics on cardiothoracicsurgery and vascular surgery which resulted in the contents of this thesis. Parallel tohis PhD he started a post-graduate training in 2006 to obtain the European degree ofQualified Medical Engineer, which was achieved in June 2008.

From November 2008 on, he is employed as medical engineer in the CatharinaHospital Eindhoven with the goal to support and enhance the clinical research inthe cardiovascular departments and to enhance the cooperation with the EindhovenUniversity of Technology.

123

124 Curriculum Vitae

Publications

Papers in peer reviewed journals

• M. van ’t Veer, N.H.J. Pijls, W.H. Aarnoudse, J.J. Koolen, and F.N. van deVosse. Evaluation of the hemodynamic characteristics of drug-eluting stentsat implantation and at follow-up. European Heart Journal, 27:1811-1817, 2006.

• M. van ’t Veer, A.H.M van Straten, F.N. van de Vosse, and N.H.J. Pijls. Influenceof orientation of bi-leaflet valve prostheses on coronary perfusion pressure inhumans. Interactive Cardiovascular and Thoracic Surgery, 6:588-592, 2007.

• W.H. Aarnoudse, M. van ’t Veer, N.H.J. Pijls, J. ter Woorst, S. Vercauteren,P.A.L. Tonino, M.C.F. Geven, M.C.M. Rutten, E. van Hagen, B. DeBruyne, andF.N. van de Vosse. Direct volumetric blood flow measurement in coronaryarteries by thermodilution Journal of the American College of Cardiology,50:2294-2304, 2007.

• M. van ’t Veer, J.Buth, M.A.G. Merkx, P.A.L. Tonino, N.H.J. Pijls, H. van denBosch, and F.N. van de Vosse. Biomechanical properties of abdominal aorticaneurysms assessed by simultaneously measured pressure and volume changesin humans. Journal of Vascular Surgery, 2008, accepted

• O. Frobert, M. van ’t Veer, W.H. Aarnoudse, U. Simonsen, J.J. Koolen, andN.H.J. Pijls. Acute Myocardial infarction and underlying stenosis severity.catheterisation and Cardiovascular Interventions, 70:958-965, 2007.

• N.H.J. Pijls, P. van Schaardenburgh, G. Manoharan, E. Boersma, J.W. Bech,M. van ’t Veer, F. Bar, J. Hoorntje, J.J. Koolen, W. Wijns, and B. DeBruyne.Percutaneous coronary intervention of functionally nonsignificant stenosis: 5-year follow-up of the DEFER study. Journal of the American College of Cardiology,49:2105-2111, 2007.

125

126 Publications

Refereed proceedings

• M. van ’t Veer, M.C.M. Rutten, J. Buth, N.H.J. Pijls, F.N. van de Vosse.Feasibility of determination of mechanical properties of abdominal aorticaneurysms by simultaneous pressure and volume measurements in-vivo . In:2008 Summer Bioengineering Conference, Marco Island, Florida, 2008.

Abstracts

• M. van ’t Veer, W.H. Aarnoudse, M.C.F. Geven, B. DeBruyne, F.N. van de Vosse,N.H.J. Pijls. Absolute blood flow measurements in selective coronary arteriesby continuous thermodilution: animal validation. European Heart Journal,28:671(suppl), 2007.

• M. van ’t Veer, G. Salvatore, A.H.M van Straten, F.N. van de Vosse, N.H.J. Pijls.Influences of orientation of bi-leaflet mechanical heart valve prostheses in aorticposition on coronary blood flow: in-vitro evaluation. Netherlands Heart Journal,15:17 (suppl 2), 2007.

hemodynamic measurements in coronary, valvular, and ... · hemodynamic measurements in coronary,...

Documents