F A C U L T Y O F H E A L T H A N D M E D I C A L S C I E N C E S

U N I V E R S I T Y O F C O P E N H A G E N

This thesis has been submitted to the Graduate School at the Faculty of

Health and Medical Sciences, University of Copenhagen

PhD thesis

Tor Biering-Sørensen

Cardiac Time Intervals by Tissue Doppler Imaging M-mode

Echocardiography: Reproducibility, Reference values, Association

with Clinical Characteristics and Prognostic Implications

Academic advisors:

Jan Skov Jensen, MD, PhD, DMSc, Professor (Faculty supervisor)

Rasmus Møgelvang, MD, PhD

1

Preface

My PhD thesis could not have been realized without the help, guidance and ideas from a great

number of people, to whom I am deeply indebted.

Firstly, I would like to thank my main advisor and mentor Jan Skov Jensen who more than six years

ago, in June 2008, included me in his research group. I started during my medical studies with a

pregraduate research-year. Although young and inexperienced, Jan had confidence in me and I was

given the opportunity being part of several of his projects. Among these were the echocardiographic

substudy of the Copenhagen City Heart Study, the “Prognosis After Percutaneous Coronary

Intervention” study and the INVASIVE study. The present PhD thesis is conducted as a part of

these studies and I am thankful to all who have contributed to these investigations. I am particularly

indebted to Jan for his tremendous scientific skills, his reliability, commitment and not least

availability when needed. I admire Jan’s work and I am grateful for his superb mentorship. We have

spent many hours discussing new projects and ideas, and I look very much forward working

together with Jan many years to come.

Special thanks also go to Rasmus Møglevang, who was one of Jan’s PhD-students when I started as

a medical student in Jan’s group. Rasmus taught me how to analyze tissue Doppler

echocardiography and introduced me to the world of advanced echocardiographic techniques. He

has, both during my pregraduate research year and during my PhD-study period, been an enormous

help. His dedication to research and thorough review and corrections of my work has been

invaluable. Throughout my PhD Rasmus has been a huge support.

Søren Galatius has throughout my PhD been a very valued partner and collaborator. He has a

unique talent to identify the key scientific messages and to guide how to communicate these to the

world. This talent of his has improved several of my manuscripts and presentations, and I am

grateful for the continuous guidance that Søren gives me.

2

You can’t say echocardiography without mentioning Peter Søgaard. Peter’s tremendous knowledge

in advanced echocardiography has during my pregraduate research and the beginning of my PhD

been an enormous support. His ideas and knowledge in cardiac mechanics has been a source of

continuous inspiration.

From the Copenhagen City Heart study I would like to thank Peter Schnohr, Merete Appleyard,

Jacob Louis Marott and Lars Måge who have assisted me whenever needed. My PhD would not

have been realized without the unique possibilities that the Copenhagen City Heart Study provides.

I’m therefore very grateful to all who have contributed to and allowed me to work with this

outstanding population study.

I would also like to thank Sune Pedersen who has done a tremendous work with the “Prognosis

After Percutaneous Coronary Intervention” study. Sune was the first person to inaugurate me to the

world of clinical research, databases and statistics. However, above all, I would like to thank Sune

for his friendship and for the many hours of laughter and good times we have spent together.

Also, I have been so lucky to experience the unique atmosphere of being a part of

“Forskningsvillaen” at Gentofte Hospital. During my PhD we were moved to new surroundings but

the atmosphere and comradeship of the “villaen” prevailed. I would like to thank all my fellow

PhD-students and all the project nurses for making my PhD time the best time of my life. Each and

every one of you has made this experience unforgettable. Special thanks goes to Magnus Thorsten

Jensen, Peter Godsk Jørgensen, Martina Chantal de Knegt and Flemming Javier Olsen for their

great contribution to my PhD, hours of discussions regarding statistics, methodology, clinical

research, advanced echocardiography, and simply for being my highly treasured friends. I would

also like to express my sincerest gratitude to my father, Fin Biering-Sørensen, for helping me to

transform my often incomprehensible way of writing into intelligible text on many occasions. I am

also grateful to the rest of my family and friends for their indispensable support.

3

Last but not least, I would like to thank my beautiful and wonderful wife, Sofie, for loving me,

supporting me and for believing in me despite the many hours of work I’ve had to prioritize to my

research for making this PhD a reality.

Tor Biering-Sørensen, Boston, USA, October 2014

4

Table of contents

1. Introduction 9

1.1 Objectives

1.2 Background

2. Methods 16

2.1 Study population (The Copenhagen City Heart Study) paper I+II+IV+V

2.2 Study population (Prognosis After Percutaneous Coronary Intervention) paper III

2.3 Study population (INVASIVE study) paper II

2.4 Conventional and novel echocardiography

2.5 Cardiac Time Intervals obtained by the conventional method (paper I & II)

2.6 Cardiac Time Intervals obtained by color TDI M-mode (paper I-V)

2.7 Follow-up (paper I & III-V)

2.8 Statistics (paper I-V)

3. Summary of main results 25

3.1 Paper I

3.2 Paper II

3.3 Paper III

3.4 Paper IV

3.5 Paper V

4. Discussion 45

4.1 Normal values of the cardiac time intervals obtained by TDI M-mode

4.2 Comparison between MPIConv and MPITDI

4.3 Cardiac time intervals obtained by TDI M-mode in hypertension

4.4 Prognostic utility of cardiac time intervals obtained by TDI M-mode

4.5 Incremental prognostic value of adding an assessment of the IVRT/ET and the

MPI obtained by TDI M-mode

5. Limitations 53

6. Conclusion 56

7. Perspectives 58

8. Summaries 60

8.1 English Summary

8.2 Dansk Resumé

9. References 66

10. Papers I-V 76

5

List of Papers

I. Biering-Sørensen T, Mogelvang R, Pedersen S, Schnohr P, Sogaard P, Jensen JS.

Usefulness of the Myocardial Performance Index Determined by Tissue Doppler Imaging

M-Mode for Predicting Mortality in the General Population. Am J Cardiol. 2011 Feb

1;107(3):478-83.

II. Biering-Sørensen T, Mogelvang R, de Knegt MC, Olsen FJ, Galatius S, Jensen JS. Cardiac

Time Intervals by Tissue Doppler Imaging M-mode: Normal Values and association with

established Echocardiographic and Invasive measures of Systolic and Diastolic function.

Submitted.

III. Biering-Sørensen T, Mogelvang R, Søgaard P, Pedersen SH, Galatius S, Jørgensen PG,

Jensen JS. Prognostic value of cardiac time intervals by tissue Doppler imaging M-mode in

patients with acute ST-segment-elevation myocardial infarction treated with primary

percutaneous coronary intervention. Circ Cardiovasc Imaging. 2013 May 1;6(3):457-65.

IV. Biering-Sørensen T, Mogelvang R, Schnohr P, Jensen JS. The Cardiac Time Intervals

measured by Tissue Doppler Imaging M-mode: Association to Hypertension, Left

Ventricular Geometry and Future Ischemic Cardiovascular Diseases. Submitted.

V. Biering-Sørensen T, Mogelvang R, Jensen JS. Prognostic value of cardiac time intervals

measured by tissue Doppler imaging M-mode in the General Population. Submitted.

6

Abbreviations

a’ = Peak late longitudinal diastolic myocardial velocity by Tissue Doppler Imaging

AMI = Acute Myocardial Infarction

A-wave = Atrial diastolic filling by pulsed-wave Doppler

AVC = Aortic Valve Closing

AVO = Aortic Valve Opening

BMI = Body Mass Index

BP = Blood Pressure

CHF = Heart Failure

CI = Confidence Interval

CV = Coefficient of variation

dP/dt max = The rate of LV pressure rise in early systole

dP/dt min = The rate of LV pressure decline in early diastole

DT = Deceleration Time of the E-wave

e’ = Peak early longitudinal diastolic myocardial velocity by Tissue Doppler Imaging

eGFR = Estimated Glomerular Filtration Rate

E-wave = Early diastolic filling by pulsed-wave Doppler

ECG = Electrocardiogram

ESH/ESC = European Society of Hypertension/European Society of Cardiology

ET = Ejection Time

GLS = Global Longitudinal Strain

GL SR s = Global Longitudinal systolic strain rate

GL SR e = Global Longitudinal early diastolic strain rate

HR = Hazard Ratio

7

ICVD = Ischemic Cardiovascular Disease

IHD = Ischemic Heart Disease

IS = previous Ischemic Stroke

IVCT = Isovolumic Contraction Time

IVRT = Isovolumic Relaxation Time

LA = Left Atrium

LAD = Left Atrium Diameter in end-systole

LV = Left Ventricle

LVEF = LV Ejection Fraction

LVH = LV Hypertrophy

LVIDd = LV diameter in end-diastole

LVMI = LV Mass Index

MACE = Major Adverse Cardiovascular Events

M-mode = Motion mode

MPI = Myocardial Performance Index

MPIConv = MPI obtained by the conventional method

MPITDI = MPI obtained by using TDI M-mode through the mitral valve

MV = Mitral Valve

MVC = Mitral Valve Closing

MVO = Mitral Valve Opening

NRI = Net Reclassification Index

PEP = Pre Ejection Period

pPCI = Primary Percutaneous Coronary Intervention

re-MI = New Myocardial Infarction

8

RWT = Relative Wall Thickness

SERCA2a = Sarco/Endoplasmic Reticulum Ca2+

-ATPase 2a

SHR = Subdistribution Hazard Ratios

STEMI = ST-Segment Elevation Myocardial Infarction

TDI = Tissue Doppler Imaging

TnI = Troponin I

9

1. Introduction

The cardiac time intervals are defined by the opening and closing of the valves of the heart, which

in turn is determined by pressure differences across the valves. All clinicians are familiar with

audible assessing the cardiac time intervals. By using their stethoscope, audible changes in the

cardiac time intervals can be assessed. Just by auscultating the first (S1) and second (S2) heart

sound the clinician can evaluate the total systolic time (isovolumic contraction time (IVCT) +

ejection time (ET)), which is the time difference between the mitral valve closing (MVC) and the

aortic valve closing (AVC). Additionally, the time difference from the S2 to the S1 is the total

diastolic time (isovolumic relaxation time (IVRT) + the diastolic filling period), which is the time

difference between the AVC and the MVC. The cardiac time intervals have therefore been

accessible for the clinician to evaluate since the beginning of the 19th

century where the stethoscope

was invented. Thus researchers have for more than a century investigated the usability of theses for

diagnostic and predictive purposes1–7

. However, the methods of assessing the cardiac time intervals

have changed significantly through the last century. From using sphygmographic recording of the

arterial pulse initiated in the 19th

century1 to using simultaneous photographic recordings of the

electrocardiogram, phonocardiogram, and the central arterial pulse4,5

in the beginning of the 20th

century, to simultaneous recordings of the electrocardiogram, the central arterial pulse, and M-mode

echocardiography in the late 20th

century6, to Doppler echocardiography in the mid-1990

th7–9, and

finally tissue Doppler imaging (TDI) velocity traces of the ventricular myocardium in the beginning

of the 21th

century10–12

. The progress in how to obtain the cardiac time intervals has made them

easily accessible to attain. Therefore, the evaluation of the usefulness of cardiac time intervals as

diagnostic and prognostic markers in the clinical setting has become increasingly meaningful.

A decade ago, yet another echocardiographic method of obtaining the cardiac time intervals

evolved. Using this method, the cardiac time intervals are obtained by evaluating the mitral valve

10

(MV) movement through the cardiac cycle using a simple color TDI M-mode analysis13,14

. This

novel echocardiographic method seemed to have several advantages compared to the conventional

echocardiographic methods of obtaining the cardiac time intervals. However, in order to assess the

usability of this method of obtaining the cardiac time intervals in clinical practice, future studies are

necessary.

1.1 Objectives

The aims of the present thesis were:

1) To compare the reproducibility, association with clinical characteristics and the usefulness

in predicting all-cause mortality in the general population, for the Myocardial Performance

Index (MPI) obtained by the conventional method of obtaining the MPI (MPIConv), as

described by Tei and colleagues7,9

and by the novel method of obtaining the MPI (MPITDI),

using TDI M-mode.

2) To define normal values of the cardiac time intervals obtained by TDI M-mode through the

MV, and to evaluate the association of the novel MPITDI and the conventional method of

obtaining the MPI (MPIConv), with established echocardiographic and invasive measures of

systolic and diastolic function.

3) To evaluate the prognostic value of the cardiac time intervals and the combined index of

systolic and diastolic performance, the MPI, obtained by TDI M-mode method in patients

with ST-Segment Elevation Myocardial Infarction (STEMI) treated with Primary

Percutaneous Coronary Intervention (pPCI).

4) To investigate if the cardiac time intervals are able to identify miniscule cardiac

impairments in individuals with hypertension, which are unrecognized by conventional

echocardiography, and if these time intervals are affected according to blood pressure (BP)

11

severity and left ventricular geometry. Additionally, to investigate if these cardiac time

intervals can be used to predict hypertension related diseases such as ischemic

cardiovascular diseases.

5) To investigate whether the combined indexes containing information on both systolic and

diastolic function (the IVRT/ET and the MPI), provide prognostic information regarding the

risk of future cardiovascular disease and cardiovascular mortality in the general population,

incremental to clinical predictors and conventional echocardiographic measures of systolic

and diastolic function.

1.2 Background

Cardiac Time Intervals at Cellular Level

The preservation of normal cardiac time intervals is intimately related to normal cardiac physiology

and function. Both contraction and relaxation of the heart is orchestrated by the recycling of

calcium in the myocardial cells. This recycling of calcium ions is initiated when the action

potential, origination from the sinus node, propagating from myocyte to myocyte through the gap

junctions, depolarizes the plasma membrane. This activates the voltage-gated L-type Ca2+

channels

in the sarcolemma. The ensuing Ca2+

influx then triggers a much greater Ca2+

release from the

sarcoplasmic reticulum via ryanodine receptors (RyR2) in a process called Ca2+

-induced Ca2+

release15

. The Ca2+

-induced Ca2+

release results in a tenfold increase in cytoplasmic Ca2+

concentrations. The Ca2+

-ions bind to troponin C which in turn results in actin-myosin cross-bridge

formation. ATP fuels the subsequent cross-bridge cycling which translocates the myosin heads

along the actin filaments, which results in the contraction of the myocyte. The contraction will

proceed as long Ca2+

(and ATP) is available in the cytosol, why relaxation is initiated by the

removal of Ca2+

. Cytosolic Ca2+

is pumped back into the sarcoplasmic reticulum by a Ca2+

pump

12

(SERCA2a)16,17

, and extruded into the extracellular fluid from the cardiomyocyte via the

sarcolemmal Na-Ca exchanger18

. This recycling of Ca2+

happens approximately once every second

and is, as long as normal cardiac physiology is maintained, orchestrated with a precision of

millisecond intervals. However, in the ailing heart, maladaptive changes result in depressed

intracellular Ca2+

cycling and lower Ca2+

concentrations in the sarcoplasmic reticulum such that any

given action potential leads to less Ca2+

influx to the cytosol and produces slower and less force

during contraction19–21

. Additionally, a reduced synchrony of sarcoplasmic reticulum Ca2+

release

and prolonged Ca2+

release in failing cardiomyocytes from human hearts has been ducumented22

.

Also the SERCA2a pump has been demonstrated to be impaired in the failing heart leading to

delayed reuptake of the Ca2+

to the sarcoplasmic reticulum causing diastolic dysfunction23–25

.

Therefore, the rate of calcium recycling and therefore also the amount of time the heart remains in

different phases of the cardiac cycle will indeed change in the progression of heart disease.

Cardiac Time Intervals at Organ Level

In the ailing myocardium, the cardiac time intervals will change during disease progression26–29

. As

left ventricular (LV) systolic function deteriorates, the time it takes for myocardial myocytes to

achieve an LV pressure equal to that of aorta increases, resulting in a prolongation of IVCT29

.

Furthermore, the ability of myocardial myocytes to maintain the LV pressure decreases, resulting in

reduction in the ET29

. As LV diastolic function declines further, early diastolic relaxation proceeds

more slowly, explaining the prolongation of IVRT. Consequently, the IVRT/ET and the MPI,

defined as (IVCT + IVRT)/ET, will detect cardiac dysfunction with an increase, irrespective of

whether the LV is suffering from impaired systolic or diastolic function30,31

. The MPI increases

even in patients with severe diastolic dysfunction with a restrictive filling pattern, which is seen as a

short IVRT time32

. This is attributable to the increase in IVCT or decrease in ET, illustrating the

13

subtle impaired systolic function identified by novel echocardiographic parameters which is found

in patients with severe diastolic dysfunction, even though they display a preserved LV ejection

fraction33,34

. Therefore, the cardiac time intervals, especially the combined indexes containing

information on both systolic and diastolic performance (the IVRT/ET and the MPI), may be useful

to identify subtle impairments in the cardiac function (both systolic and diastolic) in the general

population, which are unnoticed by conventional echocardiography35,36

.These echocardiographic

parameters may thus identify patients in high risk of future fulminant cardiovascular disease.

Methods of obtaining the cardiac time intervals

Cardiac intervals have been suggested and investigated as potential markers of cardiac dysfunction

for several decades5–7,9

. In 1964, Weissler and colleagues derived an index of systolic function

defined as the preejection period (PEP = defined as the Q-wave of the electrocardiogram – the ET

from the central artery pulse) divided by the ET (PEP/ET), using simultaneous photographic

recordings of the electrocardiogram, phonocardiogram, and the central arterial pulse5. This index

was significantly impaired in patients with heart failure compared to normal individuals, despite no

difference in electromechanical systole, defined as the time from beginning of the QRS to the end

of the T-wave. However, this index was only a measure of systolic function, why Mancini and

colleagues6 in 1982 incorporated IVRT into an index which they named the isovolumic index

(IVCT + IVRT)/ET in the attempt to capture both the systolic and diastolic function in one measure.

The sum of IVCT and IVRT was determined by subtracting the ET (obtained from the central artery

pulse) from the peak of the R wave on the electrocardiogram to the onset of mitral valve opening,

determined from M-mode echocardiography6. However, the interval from the peak of the R wave to

the onset of mitral valve opening contains an interval of electromechanical delay, which can be

pronounced in patients with left bundle branch block. However, a main concern has always been

14

that these cardiac time intervals were very difficult to obtain in a fast, easy, non-invasive and

reproducible manner.

Tei and colleagues proposed in 1995 to overcome this problem by obtaining the cardiac time

intervals from the velocity curves obtained from pulsed-Doppler echocardiography of the LV MV

inflow and outflow tract and hereby calculating the index of combined systolic and diastolic

performance, the MPI7,9

. The method suggested by Tei and colleagues unfortunately also has

several limitations. Firstly, the time intervals are obtained from two projections and at least two

cardiac cycles7,9

, and may therefore be prone to changes in the cardiac time intervals caused by

heart rate variations in-between obtaining the two projections. Secondly, the IVRT and IVCT are

not obtained when using this method, only the sum of the IVRT and the IVCT7,9

. Using TDI and the

velocity curves from the myocardium is an alternative non-invasive approach to obtain the cardiac

time intervals10–12,26,27

. With this method the time intervals can be obtained from one projection and

one cardiac cycle. However, the time intervals obtained from the TDI velocity curves are prone to

regional differences in myocardial activation, mechanics and physiology37

. These regional

differences in the cardiac time intervals will be attenuated in persons suffering from, e.g. bundle

branch blocks, ischemic heart disease or LV dyssynchrony. However, we can overcome this

problem by analysing the global time intervals through evaluating the MV movement by a simple

color TDI M-mode analysis13

, instead of the regional velocity curves. Thus, using color TDI M-

mode through the mitral leaflet to estimate the cardiac time intervals may be an improved method

reflecting global cardiac time intervals and eliminating beat-to-beat variation and regional

differences. Furthermore, the precision and reproducibility of the cardiac time intervals and MPI

seem to be improved when using the TDI M-mode method compared to the conventional method as

described by Tei and colleagues14

. Additionally, it is not possible to obtain the cardiac time

intervals from velocity curves in patients with atrial fibrillation regardless if they are obtained by

15

the conventional method as describe by Tei and colleagues7,9

or by the newer regional TDI velocity

curve method10,11

. This is due to the absence of the atrial diastolic filling wave (A-wave) and late

diastolic myocardial velocity (a’ curve) in patients with atrial fibrillation, which are needed to

obtain the time intervals by both methods. In contrast, all cardiac time intervals can be obtained by

the novel color TDI M-mode method regardless of the presence of atrial fibrillation.

Also, the time intervals achieved by clear color shifts marking minimal changes in direction in the

MV by the color TDI M-mode method are very easy to identify (Figure 2.1). In contrast the time

intervals obtained by the conventional method7,9

or the newer regional TDI velocity curve

method10,11

are assessed from velocity curves where the signals may be scattered making it hard to

accurately define the cardiac time intervals. Finally, when good imaging quality is difficult to

obtain, it is often possible to visualize the MV in the apical view, due to the perpendicular nature

between the ultrasound beam and the MV, why assessing its longitudinal movement by color TDI

M-mode nearly always is manageable.

16

2. Methods

2.1 Study population (The Copenhagen City Heart Study) paper I+II+IV+V

Within the Copenhagen City Heart Study, a longitudinal cohort study of cardiovascular disease and

risk factors, an echocardiographic sub study was performed38,39

. The present thesis includes all

participants from the fourth Copenhagen City Heart Study examination 2001-2003, who had an

echocardiographic examination, including TDI, performed. We were able to obtain the cardiac time

intervals from 1,915 participants aged 20 to 93 years. In paper I only the 1,100 first analysed

participants were included. Whether a participant underwent echocardiography as part of the fourth

Copenhagen City Heart Study examination was independent of his or her health status and other

risk factors.

Health Examination

All participants answered a self-administered questionnaire and underwent thorough physical

examinations. BP of all participants was measured with the London School of Hygiene

sphygmomanometer. Hypertension was defined as systolic BP of 140 mm Hg or above, diastolic

BP of 90 mm Hg or above, or use of antihypertensive medication40

. A 12-lead electrocardiogram

(ECG) was recorded at rest in a supine position and coded according to the Minnesota code. Plasma

cholesterol and glucose values were measured on non-fasting venous blood samples41

. Diabetes

mellitus was defined as plasma glucose concentration of 11.1 mmol/L or above, use of insulin or

other antidiabetic medicine, self-reported disease, or hemoglobin A1c level of 7.0% or above42,43

.

Ischemic heart disease (IHD) was defined as a history of hospital admission for acute coronary

artery occlusion, percutaneous coronary intervention or coronary artery bypass grafting, or major

ischemic alterations on the ECG as defined by Minnesota codes 1.1 to 3.

17

2.2 Study population (Prognosis After Percutaneous Coronary Intervention)

paper III

From September 2006 to December 2008 a total of 391 patients were admitted with a STEMI,

treated with pPCI, and underwent a detailed echocardiographic examination at Gentofte University

Hospital, Denmark. Five patients were excluded due to inadequate quality of the echocardiographic

examination in regard to obtaining the cardiac time intervals.

The definition of a STEMI was: Presence of chest pain for >30 minutes and <12 hours, persistent

ST-segment elevation ≥2 mm in at least 2 contiguous precordial ECG-leads or ≥1 mm in at least

two contiguous limb ECG-leads (or a newly developed Left Bundle Branch Block) combined with a

Troponin I (TnI) increase >0.5 μg/L.

Baseline data were collected prospectively. Hypertension was defined as use of blood pressure-

lowering drugs on admission. Diabetes was defined as fasting plasma glucose concentration ≥7

mmol/L or non-fasting plasma glucose concentration ≥11.1 mmol/L or the use of glucose-lowering

drugs on admission.

TnI was measured immediately upon admission and after 6 and 12 hours.

Primary PCI procedure: pPCI was performed according to contemporary interventional

guidelines using pre-treatment with 300 mg acetyl salicylic acid, 600 mg clopidogrel and 10,000 IU

of unfractionated heparin. Glycoprotein IIb/IIIa inhibitors were used at the discretion of the

operator. Multivessel disease was defined as 2- or 3 vessel disease and complex lesions as type C-

lesions. Subsequent medical treatment included anti-ischemic, lipid-lowering and anti-thrombotic

drugs.

18

2.3 Study population (INVASIVE study) paper II

This study population consists of patients suspected of ischemic heart disease that underwent

coronary angiography, left heart ventricle catheterization and echocardiography (n = 44) at

Department of Cardiology, Gentofte Hospital, University of Copenhagen between April 2009 and

February 2011.

Invasive Measurements: LV pressures were measured, and an echocardiogram was preformed

according to a standardized protocol. Invasively obtained pressure curves were obtained using 5

French pigtail catheters which were placed in the LV chamber and the pressure set and transducer

was calibrated. Measurements were obtained over at least three cardiac cycles and saved as hard

copies and digitalized in relation to this study. The LV pressure curves were digitalized using Dagra

Version 2.0.12.35924. In all cardiac cycles the rate of LV pressure rise in early systole (dP/dt max)

and the rate of LV pressure decline in early diastole (dP/dt min) were measured and averaged over

all the cardiac cycles.

2.4 Conventional and novel echocardiography

All echocardiograms were obtained using Vivid ultrasound systems (GE Healthcare, Horten,

Norway). All participants were examined with conventional two-dimensional echocardiography and

color TDI. All echocardiograms were stored and analyzed offline with commercially available

software (EchoPac, GE Medical, Horten, Norway).

19

In the echocardiographic sub study of the Copenhagen City Heart Study, regional function was

evaluated by the 16 standard segments model, as suggested by the American Society of

Echocardiography44

. Left ventricular ejection fraction (LVEF) was evaluated by one observer on

the basis of the wall motion index score. LV systolic dysfunction was defined as LVEF below 50%.

From the parasternal long-axis view, the LV diameter in end-diastole (LVIDd), wall thickness and

the left atrium diameter in end-systole (LAD) were measured. LV mass index (LVMI) was

calculated as the anatomic mass44

divided by body surface area45

. Pulsed wave Doppler at the

apical view was utilized to record mitral inflow between the tips of the mitral leaflets. Peak velocity

of early (E) and atrial (A) diastolic filling and deceleration time of the E-wave (DT) were measured

and the E/A-ratio was calculated.

In paper IV left ventricular hypertrophy (LVH) was defined as LVMI ≥ 96 g/m2 for women and ≥

116 g/m2 for men

44. Left ventricular dilatation was considered present if LVIDd/height ≥ 3.3

cm/m44

. Furthermore, diastolic dysfunction was defined as DT < 140 ms and E/A<50 years > 2.5,

E/A50–70 years > 2, or E/A>70 years > 1.5, respectively. Participants were stratified according to left

ventricular geometry. Normal geometry was defined as a relative wall thickness (RWT) ≤ 0.42 and

absence of LVH, concentric remodeling was defined as a RWT > 0.42 and absence of LVH,

eccentric hypertrophy was defined as a RWT ≤ 0.42 and the presence of LVH and concentric

hypertrophy was defined as a RWT > 0.42 and the presence of LVH44

. A normal conventional

echocardiographic examination was defined as the absence of LVH, dilatation, LVEF < 50%, and

diastolic dysfunction.

In the “Prognosis After Percutaneous Coronary Intervention” study and “INVASIVE” study the LV

and left atrium (LA) volumes and LVEF were obtained using modified biplane Simpson’s

method44

. Pulsed-wave TDI tracings were obtained with the range gate placed at the septal and

20

lateral mitral annular segments in the 4-chamber view. The peak longitudinal early diastolic (e’)

velocity was measured and the average was calculated from the lateral and septal velocities and

used to obtain the E/e’. Also, two-dimensional strain echocardiography was performed from the

apical 4-chamber, 2-chamber and apical long-axis view. By speckle tracking, the endocardial border

was traced in end systole. The integrity of speckle tracking was automatically detected and visually

ascertained. In case of poor tracking, the region of interest tracing was readjusted. Global

longitudinal strain (GLS) was measured in all three apical views and averaged to provide GLS.

Furthermore, in each three apical views global longitudinal systolic strain rate (GL SR s) and global

longitudinal early diastolic strain rate (GL SR e) were measured, and averaged to provide global

estimates.

Lastly, in paper III diastolic function was assessed using mitral inflow velocity profiles and pulsed-

wave TDI tracings from the septal and lateral mitral annulus46

. Patients with atrial fibrillation were

defined as having diastolic dysfunction, and were graded as 1, 2 or 3 according to the value of

E/e’47

.

2.5 Cardiac Time Intervals obtained by the conventional method (paper I & II)

In paper I & II the time intervals were measured using the method described by Tei and

colleagues7,9

. The interval from MV closing to opening was determined as the period from the end

to onset of mitral inflow (A to E time) obtained from the pulsed-wave Doppler at the apical

position. LV ET was determined as the period from onset to end of LV outflow. MPIConv was

calculated as: (A to E time – ET)/ET7,9

.

21

2.6 Cardiac Time Intervals obtained by color TDI M-mode (paper I-V)

In all the papers included in the present thesis the cardiac time intervals were obtained by the novel

color TDI M-mode method. By this method, the cardiac time intervals were obtained by placing a

2-4 cm straight M-mode line through the septal half of the MV in the color TDI 4-chamber view,

whereby the cardiac time intervals were measured directly from the color diagram13,14

(Figure 2.1).

The IVCT was defined as the time interval from the MV closure (MVC), determined by the color

shift from blue/turquoise to red at end-diastole, to the aortic valve opening (AVO) determined by

the color shift from blue to red (Figure 2.1). The ET was defined as the time interval from the AVO

to the aortic valve closing (AVC), determined by the color shift from red to blue at end systole

(Figure 1). The IVRT was defined as the time interval from the AVC to the MV opening (MVO),

determined by the color shift from red-orange to yellow (Figure 1).

22

Figure 2.1 The cardiac time intervals assessed by a color TDI M-mode line through the mitral

leaflet

Left: Four-chamber gray-scale (bottom) and color TDI (top) views in end-systole displaying the

position of the M-mode line used for measuring the cardiac time intervals. Right: Color diagram of

the TDI M-mode line through the mitral leaflet.

MVC = Mitral Valve (MV) Closing; AVO = Aortic Valve Opening; AVC = Aortic Valve Closure;

MVO = MV Opening.

23

2.7 Follow-up (paper I & III-V)

In papers investigating outcome, follow-up data on diseases were obtained from the Danish

National Board of Health’s National Patient Registry, using ICD-10 codes. Additionally, follow-up

data on mortality were collected from the National Person Identification Registry. Follow-up was

100% in all the papers in this thesis.

2.8 Statistics (paper I-V)

Proportions were compared using χ2-test, continuous Gaussian distributed variables with Student’s

t-test and Mann-Whitney test if non-Gaussian distributed.

The association between the cardiac time intervals and clinical parameters were tested by

univariable and multivariable regression analyses including significant confounders. Linearity,

variance homogeneity, and the assumption of normality were tested with plots of residuals. Trends

were analyzed by linear regression analyses and departure from linearity was assessed by

simultaneous assessment of linear and quadratic effects.

In paper III receiver operating characteristic curves were constructed for the IVCT/ET, IVRT/ET

and MPI in effort to find the optimal cut-off value with the highest sensitivity and specificity for

predicting the risk of the combined endpoint.

In paper I and III intra- and interobserver variabilities were determined by Bland-Altman plots of

repeated analyses of stored echocardiographic loops (n=25 for patients with sinus rhythm in both

paper I and III and also performed in 14 patients with atrial fibrillation in paper III).

Additionally, in paper I and III, the cumulative survival curves were established by the Kaplan–

Meier method and compared by the log-rank test. Hazards ratios were calculated by Cox

proportional hazards regression analyses. The assumptions of linearity and proportional hazards in

24

the models were tested graphically and tested based on the Schoenfeld residuals.

In paper III and V, cumulative incidence curves of future cardiovascular events were obtained by

competing risk Cox proportional hazards regression models. Subdistribution Hazard Ratios (SHR)

were calculated by competing risk Cox proportional hazards regression analyses. Additionally, for

paper V, Harrell’s c-statistics48

obtained from univariable Cox proportional hazards regression

models were calculated in order to test the prognostic performance of all the cardiac time intervals.

Harrell’s c-statistics were obtained from multivariable Cox proportional hazards regression models

including all the conventional echocardiographic parameters and compared with Harrell’s c-

statistics obtained from multivariable Cox proportional hazards regression models including all the

conventional echocardiographic parameters and the IVRT/ET or MPI, respectively. Predictive

models, using logistic regression models, predicting the risk of future cardiovascular diseases, were

constructed to assess the integrated diagnostic improvement and net reclassification index (NRI)

when adding the IVRT/ET and the MPI, to models already including significant predictors of the

cardiovascular outcomes. A p-value ≤ 0.05 in 2-sided test was considered statistically significant.

All analyses were performed with STATA Statistics/Data analysis, SE 12.0 (StataCorp,

Texas,USA).

25

3. Summary of main results

3.1 Paper I

The aim of this paper was to compare the reproducibility, association with clinical characteristics

and the usefulness in predicting all-cause mortality in the general population, for MPIConv and

MPITDI.

Cardiac function remained significantly impaired in the IHD group determined by significantly

higher MPIConv and MPITDI after multivariable adjustment for age, gender, body mass index, heart

rate, and mean arterial blood pressure (MPIConv: 0.44 (95% CI 0.41-0.48) vs. 0.39 (95% CI 0.38-

0.40), p<0.002; MPITDI: 0.56 (95% CI 0.53-0.59) vs. 0.51 (0.50-0.52), p<0.002. Besides being

associated with the presence of IHD, MPITDI increased independently with increasing age and mean

arterial blood pressure, whereas MPIConv increased independently with increasing age, mean arterial

blood pressure, decreasing body mass index and heart rate (Table 3.1).

MPITDI, but not MPIConv, remained independently associated with overall mortality after

multivariable adjustment for age, gender, body mass index, heart rate, mean arterial blood pressure,

and IHD (MPIConv, pr 0.1 increase: Hazard Ratio (HR) 1.11 (95% CI 0.96-1.27), p=0.15; MPITDI, pr

0.1 increase: HR 1.18 (95% CI 1.04-1.34), p=0.010.

For MPITDI intraobserver variability analysis showed a mean difference ± SD of 0.016 ± 0.038 and

interobserver variability analysis showed a mean difference of 0.010 ± 0.073 (Figure 3.1). For

MPIConv intraobserver variability analysis showed a mean difference of 0.013 ± 0.057 and

interobserver variability analysis showed a mean difference of 0.082 ± 0.114 (Figure 3.1).

26

Table 3.1 Multiple Regression Models with Myocardial Performance index obtained by the convention method (MPIConv) and by

the TDI M-mode method (MPITDI) as Dependent Variables.

MPIConv MPITDI

P-value P-value

Age, per 10 year

0.010 (0.003 to 0.017)

P=0.006

0.025 (0.019 to 0.031)

P<0.0001

Male Sex 0.019 (-0.001 to 0.040) P=0.06 0.004 (-0.013 to 0.021) P=0.64

Body mass index, per 5 kg/m2 -0.015 (-0.028 to -0.001) P=0.033 0.002 (-0.009 to 0.013) P=0.77

Heart Rate, per 10 beats per minute -0.015 (-0.024 to -0.005) P=0.002 0.006 (-0.001 to 0.014) P=0.11

Mean arterial blood pressure, per 10 mmHg 0.013 (0.005 to 0.021) P=0.002 0.018 (0.011 to 0.025) P<.0001

Two Multiple regression models with MPIConv and MPITDI as Dependent Variables, and the mentioned parameters as independent variables. 95 %

confidence intervals are cited in parentheses.

27

Figure 3.1 Bland-Altman plots of the Intra- and Interobserver Difference of MPITDI and

MPIConv.

Solid lines indicate mean difference and dotted lines the 95% limits of agreement.

28

3.2 Paper II

The aim of this paper was to define normal values of the cardiac time intervals obtained by TDI M-

mode through the MV, and to evaluate the association of MPITDI and MPIConv, with established

echocardiographic and invasive measures of systolic and diastolic function.

In participants from the echocardiographic sub study of the Copenhagen City Heart Study without

hypertension, diabetes, IHD, heart failure (CHF) or atrial fibrillation (n=989), normal values of the

cardiac time intervals were assessed (Table 3.2). IVRT, IVRT/ET and MPI all increased

significantly with increasing age in both genders (p<0.001 for all). IVCT, ET, IVRT/ET, and MPI

differed significantly between male and female gender (Table 3.2).

In our invasive validation population (n=44), MPITDI was significantly associated with invasive

(dP/dt max) and echocardiographic measures of systolic (LVEF, GLS and GL SR s) and diastolic

function (e’, GL SR e), whereas MPIConv was significantly associated with LVEF, e’ and GL SR e

(Table 3.3).

29

Table 3.2 Normal values of the cardiac time intervals in healthy participants stratified according to age category and gender

Women

(n=553)

Men

(n=421)

Overall

(n=553)

Age category

20 to 39

(n=150)

Age category

40 to 59

(n=252)

Age category

60 or above

(n=151)

p-value for

age category

difference

Overall

(n=421)

Age category

20 to 39

(n=101)

Age category

40 to 59

(n=210)

Age category

60 or above

(n=110)

p-value for

age category

difference

p-value for

gender

difference

IVRT (ms) 92 (20) 78 (16) 93 (16) 106 (20) <0.001 94 (20) 78 (15) 95 (16) 109 (18) <0.001 0.13

IVCT (ms) 36 (13) 32 (12) 37 (12) 38 (14) <0.001 34 (11) 35 (11) 33 (11) 36 (12) 0.07 0.012*

ET (ms) 293 (21) 291 (19) 296 (20) 289 (23) 0.001 281 (24) 284 (19) 280 (23) 279 (28) 0.34 <0.001*

IVRT/ET 0.32 (0.07) 0.27 (0.05) 0.31 (0.06) 0.37 (0.07) <0.001 0.34 (0.08) 0.28 (0.06) 0.34 (0.06) 0.40 (0.08) <0.001 <0.001*

IVCT/ET 0.12 (0.05) 0.11 (0.04) 0.13 (0.04) 0.13 (0.05) <0.001 0.12 (0.04) 0.13 (0.04) 0.12 (0.04) 0.13 (0.05) 0.10 0.61

MPI 0.44 (0.10) 0.38 (0.08) 0.44 (0.08) 0.51 (0.10) <0.001 0.46 (0.10) 0.40 (0.08) 0.46 (0.09) 0.53 (0.10) <0.001 0.002*

BMI=Body Mass Index; LVMI=Left Ventricular Mass Index; LVIDd= left ventricular dimension in end-diastole; LAD=Left atrium diameter in end-

systole; DT=deceleration time of early diastolic inflow; IVCT=Isovolumic Contraction Time; IVRT=Isovolumic Relaxation Time; ET=Ejection Time;

MPI = Myocardial Performance Index

Numbers in parenthesis indicates standard deviation. * remained statistical significantly different between the genders after multivariable adjustment for

age, BMI, heart rate, systolic and diastolic blood pressure, LVMI, LVIDd, LAD, and DT

30

Table 3.3 Regression Models with Myocardial Performance index obtained by the convention method (MPIConv) and by the TDI M-

mode method (MPITDI) as Dependent Variables (n=44)

MPITDI MPIConv

Beta (95% CI) p-value Beta (95% CI) p-value

Invasive measures of systolic and diastolic function:

dP/dt max, per 100 mmHg/sec increase -0.015 (-0.029 to -0.000) 0.047 0.011 (-0.005 to 0.0263) 0.18

dP/dt min, per 100 mmHg/sec decrease 0.012 (-0.003 to 0.027) 0.12 -0.001 (-0.017 to 0.016) 0.94

Conventional and novel echocardiographic measures of systolic and diastolic function:

LVEF, per 1 % increase -0.743 (-1.338 to -0.148) 0.016 -0.613 (-1.220 to -0.007) 0.048

e’, per 1 cm/sec decrease 0.035 (0.018 to 0.053) <0.001 0.021 (0.001 to 0.040) 0.026

E/e’, per 1 increase 0.003 (-0.007 to 0.013) 0.56 0.002 (-0.008 to 0.011) 0.74

GLS (%), per 1 % decrease 0.021 (0.010 to 0.033) 0.001 0.009 (-0.004 to 0.022) 0.16

GL SR s, per 1 sec-1

decrease 0.525 (0.306 to 0.745) <0.001 0.240 (-0.002 to 0.482) 0.051

GL SR e, per 1 sec-1

increase -0.481 (-0.618 to -0.343) <0.001 -0.361 (-0.511 to -0.210) <0.001

dP/dt max= the rate of LV pressure rise in early systole; dP/dt min= the rate of LV pressure decline in early diastole; LVEF=Left Ventricular Ejection

Fraction; E=peak transmitral early diastolic inflow velocity; e’=average peak early diastolic longitudinal mitral, GLS = Global Longitudinal Strain; GL

SR s = Global Longitudinal systolic strain rate; GL SR e = Global Longitudinal early diastolic strain rate

31

3.3 Paper III

The aim of this paper was to evaluate the prognostic value of the cardiac time intervals and MPI,

obtained by TDI M-mode method in patients with ST-Segment Elevation Myocardial Infarction

(STEMI) treated with Primary Percutaneous Coronary Intervention (pPCI).

Only the IVRT/ET and the MPI remained independent predictors of the combined outcome after

adjusting for age, gender, peak TnI, previous acute myocardial infarction (AMI), LVEF, GLS,

diastolic function grade, LVIDd/BSA and LVMI in patients with STEMI undergoing pPCI (Table

3.4). Even in patients with atrial fibrillation the MPI was significantly higher in patients with an

adverse outcome (0.45 (95% confidence interval (CI) 0.34-0.55) vs. 0.67 (95% CI 0.58-0.76),

p=0.005) (Figure 3.2). Even after adjusting for age, gender and LVEF the MPI remained

significantly higher in the group with an adverse outcome (0.44, 95% CI=(0.29-0.58) vs. 0.67, 95%

CI=(0.56-0.79), p=0.026). The intraobserver variability analysis showed a mean difference ± SD of

-0.004 ± 0.029 (Coefficient of variation (CV) 6%) for patients with sinus rhythm and 0.031 ± 0.055

(CV 9%) for patients with atrial fibrillation (Figure 3.3). The interobserver variability analysis

showed a mean difference of ± SD of -0.022 ± 0.084 (CV 16%) for patients with sinus rhythm and

0.017 ± 0.096 (CV 16%) for patients with atrial fibrillation (Figure 3.3).

32

Table 3.4 Unadjusted and adjusted Cox proportional hazards regression models depicting the

combined indexes of the cardiac time intervals as predictors of outcome

Combined

endpoint

(96 events)

Hazard Ratio

(95% CI)

P-value

CHF

(53 events)

Hazard Ratio

(95% CI)

P-value

re-MI

(25 events)

Hazard Ratio

(95% CI)

P-value

Mortality

(33 events)

Hazard Ratio

(95% CI)

P-value

Unadjusted model:

IVRT/ET per

0.1 increase

1.34 (1.14-1.57) <0.001 1.28 (1.03-1.59) 0.028 1.40 (1.04-1.89) 0.025 1.12 (0.84-1.50) 0.44

IVCT/ET per

0.1 increase

1.53 (1.26-1.86) <0.001 1.68 (1.34-2.11) <0.001 1.39 (0.94-2.08) 0.10 1.41 (0.99-2.00) 0.06

MPI per 0.1

increase

1.30 (1.16-1.46) <0.001 1.32 (1.14-1.54) <0.001 1.31 (1.05-1.63) 0.015 1.17 (0.95-1.43) 0.15

Multivariable model adjusted for age, gender, peak TnI, pre-MI, LVEF, diastolic function grade, LVMI and LVIDd/BSA:

IVRT/ET per

0.1 increase

1.29 (1.08-1.54) 0.005 1.21 (0.94-1.56) 0.15 1.18 (0.81-1.72) 0.40 1.20 (0.89-1.60) 0.23

IVCT/ET per

0.1 increase

1.42 (1.00-2.02) 0.053 1.43 (0.90-2.26) 0.13 1.34 (0.63-2.89) 0.45 1.42 (0.77-2.60) 0.26

MPI per 0.1

increase

1.26 (1.09-1.46) 0.002 1.22 (0.99-1.50) 0.06 1.17 (0.86-1.60) 0.32 1.20 (0.94-1.52) 0.15

Multivariable model adjusted for age, gender, peak TnI, pre-MI, GLS, diastolic function grade, LVMI and LVIDd/BSA:

IVRT/ET per

0.1 increase 1.30 (1.08-1.56) 0.005 1.22 (0.94-1.58) 0.14 1.26 (0.86-1.84) 0.23 1.20 (0.88-1.64) 0.26

IVCT/ET per

0.1 increase 1.43 (0.98-2.08) 0.06 1.53 (0.93-2.49) 0.09 1.54 (0.70-3.38) 0.28 1.16 (0.59-2.28) 0.66

MPI per 0.1

increase 1.27 (1.09-1.48) 0.002 1.25 (1.00-1.55) 0.046 1.25 (0.92-1.71) 0.16 1.16 (0.89-1.52) 0.26

Multivariable model adjusted for age, gender, peak TnI, pre-MI, LVEF, GLS, diastolic function grade, LVMI and LVIDd/BSA:

IVRT/ET per

0.1 increase

1.28 (1.07-1.53) 0.008 1.22 (0.94-1.57) 0.14 1.14 (0.77-1.68) 0.51 1.18 (0.87-1.61) 0.29

IVCT/ET per

0.1 increase

1.30 (0.89-1.90) 0.18 1.36 (0.83-2.24) 0.23 1.25 (0.56-2.75) 0.59 1.13 (0.58-2.20) 0.72

MPI per 0.1

increase

1.24 (1.07-1.45) 0.005 1.22 (0.98-1.51) 0.08 1.14 (0.82-1.57) 0.44 1.15 (0.88-1.49) 0.30

IVRT=Isovolumic Relaxation Time; IVCT=Isovolumic Contraction Time; ET=Ejection Time;

MPI=Myocardial Performance Index; peak TnI=Peak Troponin I; Pre-MI=Previous Acute Myocardial

Infarction; LVEF=Left Ventricular Ejection Fraction; LVMI=Left Ventricular Mass Index; LVIDd=Left

Ventricular Internal Diameter in Diastole; BSA=Body Surface Area, GLS=peak Global Longitudinal systolic

Strain; CHF=Heart Failure; re-MI= new myocardial infarction

33

Figure 3.2 MPI in patients with atrial fibrillation stratified according to outcome

The MPI for patients with atrial fibrillation divided according to outcome. Circles indicate absolute

values, dotted lines indicate means. MPI = Myocardial Performance Index.

34

Figure 3.3 Intra- and interobserver variability analysis

Bland-Altman plots of intra- and interobserver differences of myocardial performance index

assessed by color tissue Doppler imaging M-mode through the mitral leaflet for patients with sinus

rhythm (n=25) and for patients with atrial fibrillation (n=14) showing mean difference (solid lines)

and 95% limits of agreement (dotted lines).

35

3.4 Paper IV

The aim of this paper was to investigate if the cardiac time intervals were able to identify miniscule

cardiac impairments in individuals with hypertension, and if the time intervals were affected

according to BP severity and left ventricular geometry. Additionally, to investigate if cardiac time

intervals can be used to predict hypertension related diseases such as ischemic cardiovascular

diseases.

After multivariable adjustment for clinical variables the IVRT, IVRT/ET and MPI, remained

significantly impaired in persons with hypertension compared to participants without (Figure 3.4).

Furthermore, the IVRT, the IVRT/ET, and the MPI, remained significantly impaired after

multivariable adjustment when the analysis was confined to persons with a normal conventional

echocardiography (Figure 3.4). Additionally, they displayed a significant dose-response

relationship, with increasing severity of elevated BP and increasing LVMI (p<0.001 for all)(Figure

3.5). The IVRT/ET and MPI were powerful and independent predictors of future ischemic

cardiovascular disease (ICVD) especially in participants with known hypertension (Figure 3.6). In

addition, reclassification analysis demonstrated, that adding IVRT/ET or MPI, to the clinical

predictors from the Framingham Risk score49

and the SCORE risk chart50

(age, gender, cholesterol,

smoking status and systolic BP), yielded better predicting models with significant increase in the

categorical NRI of 2.27% (95% CI 0.17-4.37%, p=0.034) for IVRT/ET and 4.02% (95% CI 1.50-

6.53%, p=0.002) for the MPI, respectively. Additionally, reclassification analysis demonstrated,

that adding IVRT/ET or MPI, to a model also including the clinical predictors from the newer

European Society of Hypertension/European Society of Cardiology (ESH/ESC) risk chart51

(age,

gender, smoking status, cholesterol, diabetes, systolic BP, diastolic BP, LVH, chronic kidney

disease (defined as estimated Glomerular Filtration Rate (eGFR)≤60 mL/min/1,73 m2), IHD and

36

previous ischemic stroke (IS)), yielded better predicting models with significant increase in the

categorical NRI of 3.41% (95% CI 0.79-6.03%, p=0.011) for IVRT/ET and 3.44% (95% CI 0.39-

6.48%, p=0.027) for the MPI, respectively.

37

Figure 3.4 Systolic and diastolic function determined by the cardiac time intervals in participants with and without hypertension

after multivariable adjustment

Multivariable adjusted means are depicted for the cardiac time intervals in participants with and without hypertension for both the entire study cohort

and the subgroup with a normal conventional echocardiographic examination. Multivariable adjustment designates adjustment for age, gender, body

mass index, estimated glomerular filtration rate, heart rate, diabetes, cholesterol, atrial fibrillation, ischemic heart disease and previous ischemic stroke.

Normal echo designates the absence of left ventricular hypertrophy, dilatation and ejection fraction<50%, and severe diastolic dysfunction. Bars

indicate standard error. IVRT=Isovolumic Relaxation Time; ET=Ejection Time.

38

Figure 3.5 The association between the cardiac time intervals and LVMI and MAP

The association between the combined cardiac time intervals (IVRT/ET and MPI) and the left ventricular mass index and mean arterial

blood pressure. Participants were stratified into tertiles of the LVMI and MAP. Bars indicate standard error. LVMI=Left Ventricular Mass

Index; MAP=Mean arterial blood pressure; IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI=Myocardial Performance Index.

39

Figure 3.6 The cardiac time intervals as predictors of future ischemic cardiovascular disease

40

Depicting the subdistribution hazard ratios (SHR) obtained from univariable analysis (Figure 3.6A),

adjusted for age, gender, systolic and diastolic blood pressure (Figure 3.6B), and multivariable

(Figure 3.6C) competing risk Cox proportional hazards regression models describing the cardiac

time intervals as predictors of future hypertension related ICVD. The IVRT, IVRT/ET and the MPI

are also stratified according to hypertension status (Figure 3.6D). Multivariable adjustment

indicates adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic blood pressure,

diastolic blood pressure, diabetes, cholesterol, smoking status, atrial fibrillation, IHD, previous

ischemic stroke, LVEF<50%, diastolic dysfunction and LV hypertrophy. Depicting the SHR and

the 95% confidence intervals. BMI=Body Mass Index; eGFR= estimated Glomerular Filtration

Rate; LVEF=Left Ventricular Ejection Fraction; IVCT=Isovolumic Contraction Time;

IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI=Myocardial Performance Index.

IHD= Ischemic Heart Disease.

41

3.5 Paper V

The aim of this paper was to investigate whether the combined indexes containing information on

both systolic and diastolic function (IVRT/ET and MPI), provides prognostic information regarding

the risk of future cardiovascular disease and cardiovascular mortality in the general population,

incremental to clinical predictors and conventional echocardiographic measures of systolic and

diastolic function.

After multivariable adjustment for clinical predictors and conventional echocardiography, only the

combined indexes, including information on both the systolic and diastolic performance (IVRT/ET

and MPI), remained significant prognosticators (Table 3.5). Adding IVRT/ET or MPI to a model

already including all other echocardiographic parameters resulted in a significant increase in the

Harrell’s c-statistics (Figure 3.7). Finally, reclassification analysis when adding the combined

indexes (IVRT/ET or MPI), to clinical predictors (age, gender, body mass index (BMI), eGFR,

heart rate, hypertension, diabetes, smoking status and atrial fibrillation), yielded better predicting

models with a significant increase in the categorical NRI of 3.052% (95% CI 0.957-5.145%,

p=0.004) for the IVRT/ET and 3.058% (95% CI 0.159-5.957%, p=0.039) for the MPI, respectively.

In addition, when using continuous NRI, without using arbitrary risk categories, the predicting

models were improved even further, with an increase in the continuous NRI of 19.2% (95% CI 3.2-

33.0%) for the IVRT/ET and of 16.0% (95% CI 5.5-33.5%) for the MPI, respectively.

42

Table 3.5 Unadjusted and adjusted a competing risk Cox proportional hazards regression models depicting the cardiac time

intervals as predictors of future major adverse cardiovascular outcome (MACE)

Model 1 is adjusted for age, gender, BMI, eGFR, Heart Rate, hypertension, diabetes, smoking status, atrial fibrillation, ischemic heart disease, previous

ischemic stroke. Model 2 is adjusted for the same variables as Model 1 and for systolic dysfunction determined by LVEF < 50%, LA diameter, LVMI,

DT, E/A-ratio < 1 and E/e’.

SHR=Subdistribution Hazard Ratio; BMI=Body Mass Index; eGFR= Estimated Glomerular Filtration Rate; LVEF = Left Ventricular Ejection Fraction;

LVMI = Left Ventricular Mass Index; E = peak transmitral early diastolic inflow velocity; A = peak transmitral late diastolic inflow velocity; DT =

deceleration time of early diastolic inflow; e’ = average peak early diastolic longitudinal mitral annular velocity determined by color TDI;

IVCT=Isovolumic Contraction Time; IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI = Myocardial Performance Index.

Harrell’s

c-statistics

Univariable

SHR (95% CI)

P-value

Model 1

Multivariable

Adjustment

SHR (95% CI)

P-value

Model 2

Multivariable

Adjustment

SHR (95% CI)

P-value

IVRT per 10 ms

increase

0.62 1.20 (1.16-1.25) <0.001 1.06 (1.02-1.12) 0.004 1.07 (1.00-1.14) 0.051

IVCT per 10 ms

increase

0.57 1.19 (1.11-1.27) <0.001 1.10 (1.02-1.19) 0.010 1.07 (0.98-1.17) 0.14

ET per 10 ms

decrease

0.57 1.13 (1.08-1.18) <0.001 1.05 (0.99-1.11) 0.12 1.01 (0.94-1.07) 0.84

IVRT/ET per 0.1

increase

0.64 1.44 (1.31-1.58) <0.001 1.17 (1.08-1.27) <0.001 1.16 (1.00-1.34) 0.047

IVCT/ET per 0.1

increase

0.59 1.71 (1.46-2.02) <0.001 1.33 (1.10-1.60) 0.003 1.18 (0.96-1.45) 0.11

MPI per 0.1

increase

0.64 1.41 (1.33-1.50) <0.001 1.17 (1.09-1.25) <0.001 1.11 (1.03-1.23) 0.024

43

Figure 3.7 Incremental prognostic information by adding an assessment of the combined

cardiac time intervals (the IVRT/ET and the MPI) to the conventional echocardiographic

measures

The Harrell's c-statistic values obtained from the multivariable Cox proportional hazards regression models.

Conventional measures of systolic and diastolic function includes: LVEF<50%, E/e’, E/A-ratio, Dec time,

0.73

0.74

0.75

0.76

0.77

0.78

Conventional measuresof systolic and diastolic

function

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index +the IVRT/ET

0.7530148

0.758623

0.7667751

P=0.015

P=0.15

P=0.003

0.73

0.74

0.75

0.76

0.77

0.78

Conventional measuresof systolic and diastolic

function

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index +MPI

0.7530148

0.758623

0.7644389

P=0.029

P=0.15

P=0.010

Harr

el’s

c-s

tatistics

Harr

el’s

c-s

tatistics

44

LA dimension. Conventional measures of systolic and diastolic function, and the left ventricular mass index

includes: LVEF<50%, E/e’, E/A-ratio, Dec time, LA dimension and LVMI. LVEF = Left Ventricular

Ejection Fraction; LVMI = Left Ventricular Mass Index; E = peak transmitral early diastolic inflow velocity;

A = peak transmitral late diastolic inflow velocity; DT = deceleration time of early diastolic inflow; e’ =

average peak early diastolic longitudinal mitral annular velocity determined by color TDI; LA dimension =

Left Atrium end-systolic diameter; IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI =

Myocardial Performance Index.

45

4. Discussion

4.1 Normal values of the cardiac time intervals obtained by TDI M-mode

This thesis is the first to assess the normal values of the cardiac time intervals obtained by TDI M-

mode in a sample from the general population without cardiovascular disease and cardiovascular

risk factors. We found that LV function as assessed by IVRT, IVRT/ET, and MPI all decreased

with increasing age in both genders (Table 3.2). The decrease in LV cardiac function with

increasing age as assessed by IVCT and IVCT/ET, was statistically significant in females but not in

males (Table 3.2). ET decreased with increasing age in both genders; however this decrease was not

statistically significant (Table 3.2). Other novel echocardiographic parameters detecting subtle

myocardial impairments like two-dimensional strain echocardiography52

and TDI38

, have

previously, in accordance with our results, been demonstrated to detect miniscule LV dysfunction

with increasing age despite the absence of any cardiovascular disease or risk factor.

IVCT, ET, IVRT/ET, and MPI all differed significantly between male and female genders (Table

3.2). Gender differences exist both in novel echocardiographic parameters like two-dimensional

strain echocardiography52

and TDI38

, but also in the cardiac dimensions obtained by conventional

echocardiography44

, and in conventional measures of systolic53

and diastolic function38,44

. The

discrepancy we found, that woman in general display better cardiac function than men, appears to

be physiologically plausible and in accordance with previous studies38,44,52,53

.

4.2 Comparison between MPIConv and MPITDI

We found that the absolute values of MPIConv and MPITDI were different54

. The same result was

seen in previous studies comparing MPIConv to MPI obtained from regional TDI velocity curves of

the ventricular myocardium12,26,27

. Additionally, MPIConv was influenced by several physical

46

parameters compared to MPITDI (Table 3.1). MPITDI is based on analysis of the mitral leaflet

moving passively depending on shifts in pressure and blood flow in the left ventricle and left

atrium, whereas MPIConv is derived purely from blood flow. Hence, the difference in physical

parameters that define the two methods may explain the discrepancy between the absolute values.

Furthermore, it seems as if MPI based on simple cardiac mechanics are less confounded than MPI

based on calculations from versatile blood flow patterns around the mitral valve, which may make

the interpretation of MPITDI less complex.

Our results demonstrated that MPITDI (but not MPIConv) was significantly correlated with the

invasive measure of contractility and systolic function dP/dt max (Table 3.3). The only other

echocardiographic measure which was significantly correlated with dP/dt max was GL SR s, which

previously has been demonstrated to be the most accurate marker of myocardial contractile

function55

. Therefore, MPITDI may be an accurate marker of myocardial contractile function as well.

Furthermore, MPITDI increased significantly with worsening systolic function determined by all the

echocardiographic measures of systolic function (LVEF, GLS or GL SR s), whereas MPIConv only

increased significantly with worsening LVEF (Table 3.3). Likewise, MPITDI increased with

worsening diastolic function determined by the measures of diastolic function (e’ or GL SR e),

whereas MPIConv only increased significantly with worsening e’ and GL SR e (Table 3.3). Our

results illustrate that an ailing systolic or diastolic performance can be detected by an increasing

value of the MPI when assessed by TDI M-mode. Furthermore, MPITDI seem superior to MPIConv in

detecting an ailing systolic or diastolic performance regardless if evaluated with conventional

echocardiographic, novel echocardiographic or invasive measures.

In the present thesis we also compared the reproducibility of the MPIConv and MPITDI. We found

that the precision of the measurement of MPI improved when measured with the M-mode TDI

method compared to the conventional method (Figure 3.1). These findings are in line with a

47

previous study which showed that the time intervals obtained with the color TDI M-mode through

the mitral leaflet was an accurate and reproducible method, especially when intervals were used to

calculate MPITDI14

. In contrast to our results, Tei and colleagues previously demonstrated good

correlation between MPIConv and dP/dt max in a smaller study of patients with IHD56

. However, like

our results, several studies performed during the last decade, comparing MPI obtained by the

conventional method as described by Tei and collegues7,9

and by TDI, has demonstrated MPI

obtained by TDI to be a superior method14,27,57

, both when comparing the reproducibility14

and

when comparing their diagnostic and prognostic utilities27,57

. Likewise, we found that MPITDI, but

not MPIConv, remained independently associated with overall mortality after multivariable

adjustment. Additionally it is not possible to obtain cardiac time intervals from velocity curves in

patients with atrial fibrillation regardless if they are obtained by the conventional method7,9

or by

the newer TDI method using regional myocardial velocity curves10,11

. This is due to the absence of

the A and a’ velocity curves in patients with atrial fibrillation, which are needed to accurately obtain

the time intervals by both methods. In contrast all cardiac time intervals can easily be obtained by

the color TDI M-mode method regardless of the presence of atrial fibrillation. Thus, in patients with

atrial fibrillation suffering from a STEMI, MPITDI can differentiate between high and low risk

patients (Figure 3.2), even after adjustment for age, gender and LVEF. However, the reproducibility

seems lower in patients with atrial fibrillation than in patients with sinus rhythm (Figure 3.3).

Nevertheless the intra- and interobserver variability for MPITDI in patients with atrial fibrillation

seems superior to the variability found for the conventional method described by Tei and colleagues

in patients with sinus rhythm (Figure 3.1)54

, which provides more optimism for this new method.

4.3 Cardiac time intervals obtained by TDI M-mode in hypertension

48

In the present thesis we found that after multivariable adjustment, and when confining the analysis

to participants with a normal conventional echocardiographic examination, only IVRT, IVRT/ET,

and MPI remained significantly impaired in participants with hypertension (Figure 3.4). These

results were expected, since it is well known that patients with hypertension primarily suffer from

an impaired diastolic function and not an impaired systolic function58,59

, why only the time intervals

containing information about the diastolic function remained impaired after multivariable

adjustment.

However, the pivotal finding in Figure 3.4 was that the cardiac time intervals were capable of

identifying subtle impairments in the cardiac function in participants with hypertension, which were

unnoticed by conventional echocardiography (Figure 3.4). Therefore, the cardiac time intervals,

containing information about diastolic function, seem capable of identifying participants in the risk

of hypertension related asymptomatic organ damage of the heart, despite of a normal conventional

echocardiography.

In accordance with our results, a previous small scale study including 62 hypertensive patients and

15 controls, also found IVRT and MPI to be the only cardiac time intervals demonstrating impaired

cardiac function in patients with hypertension, despite of a normal systolic function60

.

Unfortunately, Cacciapuoti and colleagues did not evaluate the diagnostic utility of IVRT/ET in

their study60

.

We found a significant linear dose-response relationship, between increasing severity of elevated

blood pressure and incremental impairment in cardiac function determined by IVRT/ET and MPI

(Figure 3.5). Likewise, a previous study demonstrated that MPI increased with increasing severity

of hypertension according to left ventricular geometry61

. Accordingly, we found that IVRT/ET and

MPI displayed impaired cardiac function with increasing values of the LVMI (Figure 3.5) even

after multivariable adjustment. These findings of incremental impairment of the cardiac time

49

intervals (especially the IVRT/ET and MPI) with increasing BP and LVMI (Figure 3.5), are very

important, because a potential future echocardiographic parameter for risk stratification in

hypertension, should display incremental impairment in the measure with increasing BP, LVMI and

in all types of pathological left ventricular geometry, since the risk of cardiovascular diseases

increases continuously with increasing BP62

and LVH63,64

.

In addition, another central observation in our study was that IVRT and MPI were capable in

detecting miniscule cardiac dysfunction in participants with high normal BP compared to

participants with normal BP, when confining our analysis to participants without hypertension and

after adjustment for clinical variables. This is very important because it has previously been

demonstrated that persons with high normal BP have increased risk of developing future

hypertension65

and increased risk of ICVD66

. Thus, IVRT and MPI reveal impaired cardiac function

in persons with high normal BP, why these time intervals may aid in the early identification of

persons in high risk of future fulminant hypertension and ICVD.

4.4 Prognostic utility of cardiac time intervals obtained by TDI M-mode

Previous studies have already evaluated the prognostic value of the MPI either obtained by the

conventional method as proposed by Tei and colleagues7,9

or by using the regional TDI velocity

curves10,11

, in various populations, e.g., patients after acute myocardial infarction67,68

, elderly

men69,70

, patients with cardiac amyloidosis8, with idiopathic-dilated cardiomyopathy

71, with isolated

diastolic dysfunction34

, and in patients with systolic heart failure72

. The present thesis is the first to

evaluate the prognostic value of all the cardiac time intervals and the MPI assessed by TDI M-mode

through the mitral leaflet in a STEMI population treated by pPCI and a low risk general population.

50

We found that impairment in all the cardiac time intervals was individually associated with

increased risk of MACE (Table 3.5) and ICVD (Figure 3.6) in the general population. In our

smaller study of STEMI patients we found that the cardiac time intervals when evaluated separately

provided ambiguous and not easily comprehensible prognostic information. However when we

corrected the IVCT and the IVRT for heart rate by dividing them with ET30

, both of the combined

indexes provided incremental prognostic information with increasing tertile73

. This result is

interesting since neither the IVRT nor the ET provided incremental prognostic information when

evaluating them separately, but when combing the information about the systolic and the diastolic

performance in one index (and thereby also indirectly adjusting for heart rate), the prognostic

information becomes evident73

. However, after multivariable adjustment for other clinical and

echocardiographic predictors, only the combined indexes containing information on both the

systolic and diastolic performance (IVRT/ET and MPI) remained independent predictors of future

cardiovascular outcomes regardless which population and outcomes were examined (Table 3.4, 3.5

and Figure 3.6). The superiority of the IVRT/ET and the MPI compared to the remaining cardiac

time intervals in predicting outcomes, regardless of population, may reflect that they detect

myocardial dysfunction, irrespective of myocardial sufferings from an ailing systolic or diastolic

function73

. In contrast, the remaining cardiac time intervals only detect isolated diastolic (IVRT) or

systolic (IVCT, ET and IVCT/ET) dysfunction, respectively. Similarly, when using TDI velocities

in predicting outcome, studies have demonstrated that when combining the information on systolic

(s’) and diastolic function (e’ and a’) in one risk stratification strategy, the prognostic information

obtained improves39,74,75

. These and our results emphasize the prognostic information gained by

combining the interdependent and coherent relations of systolic and diastolic function in a simple

and feasible index.

51

4.5 Incremental prognostic value of adding an assessment of the IVRT/ET and

the MPI obtained by TDI M-mode

We found that the combined indexes containing information on both the systolic and diastolic

performance (IVRT/ET and MPI) remained not only independent predictors of future MACE in the

general population, but provided incremental prognostic value to all other echocardiographic

predictors of MACE (Figure 3.7). In addition, reclassification analysis, using arbitrary risk

categories, when adding the combined indexes (IVRT/ET or MPI) to our clinical predictors, yielded

better predicting models with significant increases in the categorical NRI of approximately 3.1% for

both IVRT/ET and MPI. In addition, when using continuous NRI, without using arbitrary risk

categories, the predicting models were improved even further, corresponding with an increase in the

continuous NRI of 19.2% for the IVRT/ET and of 16.0% for the MPI, respectively. In comparison,

neither of the conventional echocardiographic parameters evaluating systolic or diastolic function

(LVEF<50%, E/e’, LAD or E/A-ratio) or the LVMI yielded better predicting models when added to

our clinical predictors. Additionally, when evaluating the risk of ICVD, adding IVRT/ET or MPI,

to the clinical predictors from the Framingham Risk Score49

, the SCORE risk chart50

and the

ESH/ESC risk chart51

yielded better predicting models with significant increases in the categorical

NRIs. However, we found that hypertension significantly modified the relationship between all the

cardiac time intervals containing information about diastolic function (IVRT, IVRT/ET and MPI)

and the risk of future ICVD. These cardiac time intervals primarily predicted hypertension related

ICVD in participants with hypertension but not in participants without hypertension (Figure 3.6).

Similarly, a previous study demonstrated that the presence of diastolic dysfunction, determined by a

short deceleration time, only predicted all-cause mortality in hypertensive heart failure patients and

not in heart failure patients without hypertension76

. Therefore, in participants suffering from

52

hypertension, who are in particularly high risk of future ICVD77

, the IVRT/ET and MPI may be

useful cardiac measures for risk stratification. They are simple to obtain, have high reproducibility

(Figure 3.1 and 3.3)14

and encompasses information on both BP severity, left ventricular geometry

and risk of future ICVD. The superiority of the IVRT/ET and the MPI to all the conventional

echocardiographic parameters evaluating systolic or diastolic function, may again reflect that they

detect very miniscule impairment in the myocardial function, irrespective if the myocardium suffers

from an ailing systolic or diastolic function73

. Therefore, the IVRT/ET and MPI identify subtle

impairments in the cardiac function (both systolic and diastolic), that may be unnoticed by

conventional echocardiography35,36

, and can improve risk stratification strategies evaluating future

risk of fulminant cardiovascular disease over and above the traditional predictors, in the low risk

general population. In contrast, it seems that adding an assessment of the conventional

echocardiographic parameters to the clinical risk factors, does not improve the risk stratification

strategy to identify high risk people in the general population.

53

5. Limitations

Certain limitations have to be taken into account in relation to the reported results.

In paper I and III we did not investigate the cause of death. However, we assume that individuals

with cardiac dysfunction determined by echocardiography are more likely to die due to

cardiovascular causes and that, if we were able to limit our analysis to cardiovascular deaths in

paper I and III, the prognostic impact of MPI (and the other time intervals) would be even greater.

This assumption is supported by our results found in paper V where cardiac mortality was included

in the combined endpoint of MACE.

In paper III and in the INVASIVE study (part of paper II) I was not blinded for other

echocardiographic parameters, in the sense that I also measured other echocardiographic parameters

when measuring the cardiac time intervals, which theoretically could be a potential for bias.

Nevertheless the combined indexes provide independent prognostic information after adjustment for

all other systolic and diastolic parameters in our study population of paper III, which would not be

the case if the combined indexes only contain prognostic information gained from other

echocardiographic parameters. In addition, this potential bias should affect the association between

MPITDI and MPIConv with established echocardiographic and invasive measures of systolic and

diastolic function equally in the INVASIVE population, and cannot explain why the MPITDI, but not

MPIConv, is associated with most invasive and established echocardiographic measures of systolic

and diastolic function. Furthermore, we were blinded to other echocardiographic parameters, in the

sense that we did not measure other echocardiographic parameters but only TDI velocities and the

cardiac time intervals in the echocardiographic sub study of the Copenhagen City Heart Study.

In paper III echocardiography was performed median 2 days after pPCI, which was the typical

timespan for echocardiographic risk-assessment at our institution, where various degree of

myocardial stunning might be present and may impact the cardiac time intervals. However

54

myocardial function can be affected for as long as 2 weeks after the intervention78

, at which point

the myocardial function could be influenced by long-term compensatory mechanisms and the

effects of additional therapeutic interventions, so the optimal time for risk assessment after pPCI is

to be investigated.

In the echocardiographic sub study of the Copenhagen City Heart study (paper I, II, IV and V),

pulsed-wave TDI, pulmonary venous flow or Valsalva maneuver were not performed. In paper IV

we therefore only defined diastolic dysfunction as the presence of severe diastolic dysfunction as

determined from the DT and E/A-ratio79

. Participants with milder degrees of diastolic dysfunction

might have been overlooked in paper IV where a normal conventional echocardiography was

defined as the absence of LVH, dilatation and ejection fraction<50%, and severe diastolic

dysfunction. However, it is unlikely that many of our participants had mild or pseudonormal

diastolic dysfunction without concomitant structural heart disease or reduced LVEF. In addition, in

paper III (the “Prognosis After Percutaneous Coronary Intervention” population) where pulsed-

wave TDI was performed and we therefore were able to grade the diastolic function, we found that

the time intervals provided independent prognostic information after adjustment for diastolic

function grade.

Our study population undergoing heart catheterization and invasive measures of systolic and

diastolic function in paper II was small, only comprising 44 individuals, which limits our study.

However, the previous studies evaluating the association of the cardiac time intervals with invasive

measures of systolic and diastolic function have included a lower number of individuals56,80

, which

makes the validity of our results greater.

Lastly, results reported in this thesis do not demonstrate causal mechanisms, but only associations.

By adjusting for known confounders we tried to scrutinize our results. Additionally, since this thesis

55

only includes prospective non-randomized observational studies, the risk of residual confounders

always exists.

56

6. Conclusion

Based on the findings in the present thesis the following conclusions are drawn:

1) MPITDI is a measure of combined cardiac time intervals with superior reproducibility, less

confounded by association with clinical characteristics, and provides superior prognostic

information in a low-risk population when compared to MPIConv.

2) The normal values of the cardiac time intervals differ between genders, displaying that

women in general exhibit better cardiac function. Furthermore, they deteriorated with

increasing age. The MPITDI (but not MPIConv) is significantly associated with most invasive

and established echocardiographic measures of systolic and diastolic function.

3) The combined cardiac time intervals which include information on both the systolic and

diastolic function in one index (IVRT/ET and MPI) provides independent prognostic

information, regardless of rhythm, incremental to conventional and novel echocardiographic

parameters of systolic and diastolic function in patients with STEMI treated with pPCI.

4) The cardiac time intervals display a significant dose-response relationship, with increasing severity

of elevated BP and increasing left ventricular mass index. Furthermore, they identify impaired

cardiac function in participants with hypertension, not only independently of conventional risk

factors, but also in participants with a normal conventional echocardiographic examination.

Additionally, in the general population, the IVRT/ET and MPI, are powerful and independent

predictors of future hypertension related ICVD especially in participants with known hypertension.

They provide prognostic information incremental to clinical variables from the Framingham Risk

Score, the SCORE risk chart and the ESH/ESC risk chart.

5) In the general population, the combined cardiac time intervals which include information on

both the systolic and diastolic function in one index (IVRT/ET and MPI) are powerful and

independent predictors of future major cardiovascular events. In addition, they provide

57

prognostic information incremental to clinical variables and conventional echocardiographic

measures of systolic and diastolic function.

58

7. Perspectives

The results in the present thesis demonstrate that a novel, simple and reproducible method of

obtaining the cardiac time intervals and the index of combined systolic and diastolic function, the

MPI, may be useful to identify high risk persons both in the general population and in patients

suffering a STEMI. Additionally, this novel method seem superior to the conventional method of

obtaining the cardiac time intervals. These results motivate to include an assessment of the cardiac

time intervals by TDI M-mode when preforming echocardiography. However, our results do not

indicate that an assessment of the cardiac time intervals should replace conventional

echocardiography, but rather that the time intervals provide additional information, incremental to

the information we can gain from conventional echocardiography.

Several results from the present thesis need to be investigated further. Firstly, we are the first to

investigate this method, why additional validation in other populations is necessary. We found that

it was possible to obtaining the cardiac time intervals and the MPI with reasonable reproducibility

in patients with atrial fibrillation. Furthermore, it seems to be a prognostic marker in this patient

category. The potential to use this index in patients with atrial fibrillation is promising and has to be

investigated further in prospective studies. Especially in this patient category where heart rate

variability is present and often limits the usability of the conventional echocardiographic measures,

the MPI, which is corrected for heart rate by the division with ET30

, might be a fast, accurate and

superior measure of systolic and diastolic function in one index.

The association between ailing cardiac time intervals and adverse outcome also has to be

investigated in intervention studies. Does intervention improve the cardiac time intervals and if so

do these improvements result in a significant better prognosis.

59

It seems as if the cardiac time intervals (IVRT/ET and MPI) are especially useful in patients with

hypertension. It would be of interest to test the usability of the time intervals to risk stratify patients

in an actual hypertension study.

In general, especially the prognostic capabilities of the MPI obtained by TDI M-mode would be of

interest to validate in various patient populations, which might pave the way for the cardiac time

intervals to be used in the clinical setting in the future?

60

8. Summaries

8.1 English Summary

Background: The preservation of normal cardiac time intervals is intimately related to normal

cardiac physiology and function. In the ailing myocardium, the cardiac time intervals will change

during disease progression. As left ventricular (LV) systolic function deteriorates, the time it takes

for myocardial myocytes to achieve an LV pressure equal to that of aorta increases, resulting in a

prolongation of the isovolumic contraction time (IVCT). Furthermore, the ability of myocardial

myocytes to maintain the LV pressure decreases, resulting in reduction in the ejection time (ET). As

LV diastolic function declines further, early diastolic relaxation proceeds more slowly, explaining

the prolongation of isovolumic relaxation time (IVRT). Consequently, the IVRT/ET and the

myocardial performance index (MPI), defined as (IVCT + IVRT)/ET, will detect cardiac

dysfunction with an increase, irrespective of whether the LV is suffering from impaired systolic or

diastolic function. A novel method of evaluating the cardiac time intervals has recently evolved.

Using Tissue Doppler Imaging (TDI) M-mode through the mitral valve (MV) to estimate the

cardiac time intervals may be an improved method reflecting global cardiac time intervals and

eliminating beat-to-beat variation and regional differences. However, little is known about the

usability of the cardiac time intervals obtained by this novel method.

Objective: To compare the reproducibility, association with clinical characteristics, association

with established echocardiographic and invasive measures of systolic and diastolic function, and the

usefulness in predicting all-cause mortality in the general population, for the MPI obtained by the

conventional method (MPIConv) and by the novel method of obtaining the MPI (MPITDI).

Furthermore, to define normal values of the cardiac time intervals obtained by TDI M-mode

through the MV. In addition, to investigate if the cardiac time intervals are able to identify

miniscule cardiac impairments in individuals with hypertension, which are unrecognized by

61

conventional echocardiography. Lastly, to evaluate the prognostic value of the cardiac time

intervals and the combined indexes of systolic and diastolic performance, IVRT/ET and MPI,

obtained by TDI M-mode method in low risk participants from the general population and in high

risk patients with ST-Segment Elevation Myocardial Infarction (STEMI) treated with Primary

Percutaneous Coronary Intervention (pPCI).

Method: The research involved three prospective observational studies. Within the Copenhagen

City Heart Study, a large community based study of cardiovascular risk factors, the cardiac time

intervals were obtained by TDI M-mode in 1,915 participants. Additionally, in the present thesis the

cardiac time intervals were also obtained in 391 patients who were admitted with a STEMI and

treated with pPCI at Gentofte Hospital. All patients were examined by echocardiography median 2

days (IQR 1-3) after the STEMI. Lastly, we also included a population (n=44) of patients who

underwent left heart ventricular catheterization and had the MPITDI and MPIConv measured. In all

cohorts baseline data and an assessment of cardiac function were obtained by conventional

echocardiography. Follow-up data on admission were obtained from the Danish National Board of

Health’s National Patient Registry and from the national Danish Causes of Death Registry.

Results: MPITDI had superior reproducibility, was less confounded by association with clinical

characteristics, and provided superior prognostic information in a low-risk population as compared

to MPIConv. Additionally, the MPITDI (but not MPIConv) was significantly associated with most

invasive and established echocardiographic measures of systolic and diastolic function.

The cardiac time intervals displayed a significant dose-response relationship, with increasing

severity of elevated blood pressure and increasing left ventricular mass index. They identified

impaired cardiac function in participants with hypertension, not only independent of conventional

risk factors, but also in participants with a normal conventional echocardiographic examination. The

combined cardiac time intervals which include information on both the systolic and diastolic

62

function in one index (IVRT/ET and MPI) provided independent prognostic information, regardless

of rhythm, incremental to conventional and novel echocardiographic parameters of systolic and

diastolic function both in the general population and in patients with STEMI treated with pPCI.

Conclusion: The novel TDI M-mode method is superior to the conventional method of obtaining

the MPI, both when comparing reproducibility, validity, prognostic power and the ability to display

both the systolic and diastolic function in one index. The cardiac time intervals obtained by TDI M-

mode reveal impaired cardiac function in persons with hypertension even when conventional

echocardiography gives the impression of normal cardiac function. The combined cardiac time

intervals which include information on both the systolic and diastolic function in one index

(IVRT/ET and MPI) provide prognostic information incremental to conventional and novel

echocardiographic parameters of systolic and diastolic function both in the general population and

in patients with STEMI treated with pPCI.

8.2 Dansk Resumé

Baggrund: Bevarelsen af hjertets tidsintervaller inden for normalområdet er nært relateret til

normal hjertefunktion. I den syge hjertemuskel vil hjertets tidsintervaller ændres under

sygdomsprogression. Når venstre ventrikels (LVs) systoliske funktion forringes, vil den tid det

tager for hjertets muskelceller at opbygge et tryk i LV, der svarer til det der er i aorta, stige, hvilket

resulterer i en forlængelse af den isovolumetriske kontraktions tid (IVCT). Endvidere kan

muskelcellerne ikke opretholde dette høje LV tryk i så lang tid, hvilket resulterer i en reduktion af

uddrivningstiden (ET). Når LVs diastoliske funktion forringes, vil den tidlige diastolisk afslapning

forløbe langsommere, hvorved den isovolumetriske relaksations tid (IVRT) forlænges. Derfor vil

IVRT/ET og myocardial perfromance index (MPI), der er defineret som (IVRT + IVCT)/ET, kunne

63

identificere nedsat hjerte funktion med en stigning, uanset om LV lider af nedsat systolisk eller

diastolisk funktion. En ny metode til vurdering af hjertets tidsintervaller er for nylig blevet udviklet.

Hjertets tidsintervaller opmålt ved hjælp af vævs Doppler ekkokardiografi (TDI) M-mode gennem

mitralklappen (MV) kan være en forbedret metode, der afspejler hjertets globale tidsintervaller og

er upåvirket af variationen mellem de enkelte hjerteslag samt de regionale forskelle i hjertet. Men

vores viden er begrænset når det gælder anvendeligheden af hjertets tidsintervaller målt ved denne

nye fremgangsmåde.

Formål: At sammenligne reproducerbarheden, forholdet til kliniske karakteristika og etablerede

ekkokardiografiske og invasive mål for systolisk og diastolisk funktion, samt brugbarheden til at

forudsige dødelighed i den almindelige befolkning, for MPI opnået ved den konventionelle metode

(MPIConv) og ved den nye metode MPITDI. Desuden, at definere normalværdier for hjertets

tidsintervaller opmålt ved hjælp af TDI M-mode gennem MV. Hertil kommer, at undersøge om

hjertets tidsintervaller er i stand til at identificere tidlige tegn på et lidende hjerte hos personer med

forhøjet blodtryk, som ikke erkendes ved en konventionel ekkokardiografi. Slutteligt, at vurdere den

prognostiske værdi af hjertets tidsintervaller og de sammensatte indices der indeholder information

om både den systolisk og diastolisk funktion, IVRT/ET og MPI, opmålt ved hjælp af TDI M-mode

metoden i lavrisiko personer fra normalbefolkningen og hos højrisikopatienter med ST-segment

elevations myokardieinfarkt (STEMI), behandlet med primær perkutan koronar intervention (pPCI).

Metode: Afhandlingen bygger på tre prospektive observationelle studier. I Østerbroundersøgelsen,

et stort kohorte studie af kardiovaskulære risikofaktorer i normalbefolkningen, blev hjertets

tidsintervaller opmålt ved TDI M-mode i 1.915 deltagere. Derudover, blev der i forbindelse med

denne afhandling målt på hjertets tidsintervaller i 391 patienter, der blev indlagt med STEMI og

behandlet med pPCI på Gentofte Hospital. Alle patienter blev undersøgt med ekkokardiografi

median 2 dage (IQR 1-3) efter STEMI. Slutteligt, inkluderede vi en gruppe på 44 patienter, som

64

gennemgik hjerte kateterisation af LV og de fik bestemt MPITDI og MPIConv. I alle studier blev

basisdata indhentet og en konventionel ekkokardiografi udført. Opfølgningsdata blev indhentet fra

Sundhedsstyrelsens Landspatientregister og fra Dødsårsagsregisteret.

Resultater: MPITDI var MPIConv overlegen når det drejede sig om reproducerbarhed, påvirkelighed

af kliniske karakteristika, og når det drejede sig om prognostisk information i en lav-risiko

befolkning. Derudover var MPITDI (men ikke MPIConv) signifikant associeret med de fleste invasive

og etablerede ekkokardiografiske mål for systolisk og diastolisk funktion.

Hjertets tidsintervaller viste en signifikant dosis-respons-sammenhæng, med stigende sværhedsgrad

af forhøjet blodtryk og øget LV masse indeks. Disse tidsintervaller kunne identificere nedsat

hjertefunktion i deltagere med forhøjet blodtryk, ikke alene uafhængigt af traditionelle

risikofaktorer, men også i deltagere med en normal konventionel ekkokardiografisk undersøgelse.

De kombinerede hjerte tidsintervaller, som indeholder oplysninger om både den systoliske og

diastoliske funktion i et indeks (IVRT/ET og MPI) bidrog med uafhængig prognostisk information,

uanset hjerterytme, og med yderligere information i forhold til de konventionelle og nyere

ekkokardiografiske parametre vedrørende systolisk og diastolisk funktion, og dette var tilfældet

både i en lav risiko befolkning og hos patienter med STEMI behandlet med pPCI.

Konklusion: Den nye TDI M-mode metode er overlegen i forhold til den konventionelle metode til

at bestemme MPI, både når man sammenligner reproducerbarhed, validitet, prognostiske styrke og

evnen til at vise både den systoliske og diastoliske funktion i et indeks. Hjerte tidsintervaller opmålt

ved TDI M-mode metoden afslører nedsat hjertefunktion hos personer med forhøjet blodtryk, selv

når den konventionelle ekkokardiografi giver indtryk af normal hjertefunktion. De kombinerede

tidsintervaller, som indeholder information om både den systoliske og diastoliske funktion i et

indeks (IVRT/ET og MPI) indeholder prognostisk information der bidrager med yderligere

information end den vi kan opnå fra de konventionelle og nyere ekkokardiografiske parametre for

65

systolisk og diastolisk funktion, og dette gælder både i en lav risiko almindelige befolkning og i høj

risiko patienter med STEMI behandlet med pPCI.

66

9. References

1. Garrod, A. H. On some points connected with circulation of the blood arrived at from a study of

the sphygmograph. Proc Roy Soc Lond. 23, 140 (1874).

2. Bowen, W. P. Changes in heart rate, blood pressure and duration of systole from bicycling.

Amer J Physiol 11, 59 (1904).

3. Lombard, W. P. & Cope, O. M. Duration of the systole of the left ventricle of man. Amer J

Physiol 77, 263 (1926).

4. Katz, L. N. & Feil, H. S. Clinicalobservations on the dynamics of ventricular systole: I.

Auricular fibrillation. Arch Intern Med 32, 672 (1923).

5. Weissler, A. M., Harris, W. S. & Schoenfeld, C. D. Systolic time intervals in heart failure in

man. Circulation 37, 149–159 (1968).

6. Mancini, G. B. et al. The isovolumic index: a new noninvasive approach to the assessment of

left ventricular function in man. Am. J. Cardiol. 50, 1401–1408 (1982).

7. Tei, C. New non-invasive index for combined systolic and diastolic ventricular function. J.

Cardiol. 26, 135–136 (1995).

8. Tei, C. et al. Doppler index combining systolic and diastolic myocardial performance: clinical

value in cardiac amyloidosis. J. Am. Coll. Cardiol. 28, 658–664 (1996).

9. Tei, C. et al. New index of combined systolic and diastolic myocardial performance: a simple

and reproducible measure of cardiac function--a study in normals and dilated cardiomyopathy.

J. Cardiol. 26, 357–366 (1995).

10. Tekten, T., Onbasili, A. O., Ceyhan, C., Unal, S. & Discigil, B. Novel approach to measure

myocardial performance index: pulsed-wave tissue Doppler echocardiography. Echocardiogr.

Mt. Kisco N 20, 503–510 (2003).

67

11. Tekten, T., Onbasili, A. O., Ceyhan, C., Unal, S. & Discigil, B. Value of measuring myocardial

performance index by tissue Doppler echocardiography in normal and diseased heart. Jpn.

Heart J. 44, 403–416 (2003).

12. Harada, K., Tamura, M., Toyono, M., Oyama, K. & Takada, G. Assessment of global left

ventricular function by tissue Doppler imaging. Am. J. Cardiol. 88, 927–932, A9 (2001).

13. Voigt, J.-U. et al. Incidence and characteristics of segmental postsystolic longitudinal

shortening in normal, acutely ischemic, and scarred myocardium. J. Am. Soc. Echocardiogr.

Off. Publ. Am. Soc. Echocardiogr. 16, 415–423 (2003).

14. Kjaergaard, J. et al. Measurement of cardiac time intervals by Doppler tissue M-mode imaging

of the anterior mitral leaflet. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 18,

1058–1065 (2005).

15. Fabiato, A. Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum. Am. J.

Physiol. 245, C1–14 (1983).

16. Lytton, J., Westlin, M. & Hanley, M. R. Thapsigargin inhibits the sarcoplasmic or endoplasmic

reticulum Ca-ATPase family of calcium pumps. J. Biol. Chem. 266, 17067–17071 (1991).

17. Koss, K. L., Grupp, I. L. & Kranias, E. G. The relative phospholamban and SERCA2 ratio: a

critical determinant of myocardial contractility. Basic Res. Cardiol. 92 Suppl 1, 17–24 (1997).

18. Bers, D. M. & Bridge, J. H. Relaxation of rabbit ventricular muscle by Na-Ca exchange and

sarcoplasmic reticulum calcium pump. Ryanodine and voltage sensitivity. Circ. Res. 65, 334–

342 (1989).

19. Gómez, A. M. et al. Defective excitation-contraction coupling in experimental cardiac

hypertrophy and heart failure. Science 276, 800–806 (1997).

68

20. Beuckelmann, D. J., Näbauer, M. & Erdmann, E. Intracellular calcium handling in isolated

ventricular myocytes from patients with terminal heart failure. Circulation 85, 1046–1055

(1992).

21. Gwathmey, J. K. et al. Abnormal intracellular calcium handling in myocardium from patients

with end-stage heart failure. Circ. Res. 61, 70–76 (1987).

22. Louch, W. E. et al. Reduced synchrony of Ca2+ release with loss of T-tubules-a comparison to

Ca2+ release in human failing cardiomyocytes. Cardiovasc. Res. 62, 63–73 (2004).

23. Schlotthauer, K. et al. Frequency-dependent changes in contribution of SR Ca2+ to Ca2+

transients in failing human myocardium assessed with ryanodine. J. Mol. Cell. Cardiol. 30,

1285–1294 (1998).

24. Pieske, B., Maier, L. S., Bers, D. M. & Hasenfuss, G. Ca2+ handling and sarcoplasmic

reticulum Ca2+ content in isolated failing and nonfailing human myocardium. Circ. Res. 85,

38–46 (1999).

25. Hasenfuss, G. et al. Relationship between Na+-Ca2+-exchanger protein levels and diastolic

function of failing human myocardium. Circulation 99, 641–648 (1999).

26. Gaibazzi, N., Petrucci, N. & Ziacchi, V. Left ventricle myocardial performance index derived

either by conventional method or mitral annulus tissue-Doppler: a comparison study in healthy

subjects and subjects with heart failure. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc.

Echocardiogr. 18, 1270–1276 (2005).

27. Duzenli, M. A. et al. Comparison of myocardial performance index obtained either by

conventional echocardiography or tissue Doppler echocardiography in healthy subjects and

patients with heart failure. Heart Vessels 24, 8–15 (2009).

28. Bruch, C. et al. Tei-index in patients with mild-to-moderate congestive heart failure. Eur. Heart

J. 21, 1888–1895 (2000).

69

29. Carluccio, E. et al. Improvement of myocardial performance (Tei) index closely reflects

intrinsic improvement of cardiac function: assessment in revascularized hibernating

myocardium. Echocardiogr. Mt. Kisco N 29, 298–306 (2012).

30. Su, H.-M. et al. Correlation of Tei index obtained from tissue Doppler echocardiography with

invasive measurements of left ventricular performance. Echocardiogr. Mt. Kisco N 24, 252–257

(2007).

31. Uemura, K. et al. Myocardial performance index is sensitive to changes in cardiac contractility,

but is also affected by vascular load condition. Conf. Proc. Annu. Int. Conf. IEEE Eng. Med.

Biol. Soc. IEEE Eng. Med. Biol. Soc. Conf. 2013, 695–698 (2013).

32. Su, H.-M. et al. Differentiation of left ventricular diastolic dysfunction, identification of

pseudonormal/restrictive mitral inflow pattern and determination of left ventricular filling

pressure by Tei index obtained from tissue Doppler echocardiography. Echocardiogr. Mt. Kisco

N 23, 287–294 (2006).

33. Kraigher-Krainer, E. et al. Impaired systolic function by strain imaging in heart failure with

preserved ejection fraction. J. Am. Coll. Cardiol. 63, 447–456 (2014).

34. Kim, H. et al. Usefulness of tissue Doppler imaging-myocardial performance index in the

evaluation of diastolic dysfunction and heart failure with preserved ejection fraction. Clin.

Cardiol. 34, 494–499 (2011).

35. Ruggiero, A. et al. Myocardial performance index and biochemical markers for early detection

of doxorubicin-induced cardiotoxicity in children with acute lymphoblastic leukaemia. Int. J.

Clin. Oncol. 18, 927–933 (2013).

36. Varol, E. et al. Influence of obstructive sleep apnea on left ventricular mass and global function:

sleep apnea and myocardial performance index. Heart Vessels 25, 400–404 (2010).

70

37. Cui, W., Roberson, D. A., Chen, Z., Madronero, L. F. & Cuneo, B. F. Systolic and diastolic

time intervals measured from Doppler tissue imaging: normal values and Z-score tables, and

effects of age, heart rate, and body surface area. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc.

Echocardiogr. 21, 361–370 (2008).

38. Mogelvang, R. et al. Tissue Doppler echocardiography in persons with hypertension, diabetes,

or ischaemic heart disease: the Copenhagen City Heart Study. Eur. Heart J. 30, 731–739

(2009).

39. Mogelvang, R. et al. Cardiac dysfunction assessed by echocardiographic tissue Doppler

imaging is an independent predictor of mortality in the general population. Circulation 119,

2679–2685 (2009).

40. Chobanian, A. V. et al. Seventh report of the Joint National Committee on Prevention,

Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 42, 1206–1252

(2003).

41. Langsted, A., Freiberg, J. J. & Nordestgaard, B. G. Fasting and nonfasting lipid levels:

influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk

prediction. Circulation 118, 2047–2056 (2008).

42. Peters, A. L., Davidson, M. B., Schriger, D. L. & Hasselblad, V. A clinical approach for the

diagnosis of diabetes mellitus: an analysis using glycosylated hemoglobin levels. Meta-analysis

Research Group on the Diagnosis of Diabetes Using Glycated Hemoglobin Levels. JAMA J.

Am. Med. Assoc. 276, 1246–1252 (1996).

43. Rowley, K. G., Daniel, M. & O’Dea, K. Screening for diabetes in Indigenous populations using

glycated haemoglobin: sensitivity, specificity, post-test likelihood and risk of disease. Diabet.

Med. J. Br. Diabet. Assoc. 22, 833–839 (2005).

71

44. Lang, R. M. et al. Recommendations for chamber quantification: a report from the American

Society of Echocardiography’s Guidelines and Standards Committee and the Chamber

Quantification Writing Group, developed in conjunction with the European Association of

Echocardiography, a branch of the European Society of Cardiology. J. Am. Soc. Echocardiogr.

Off. Publ. Am. Soc. Echocardiogr. 18, 1440–1463 (2005).

45. Du Bois, D. & Du Bois, E. F. A formula to estimate the approximate surface area if height and

weight be known. 1916. Nutr. Burbank Los Angel. Cty. Calif 5, 303–311; discussion 312–313

(1989).

46. Nagueh, S. F. et al. Recommendations for the evaluation of left ventricular diastolic function by

echocardiography. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 22, 107–133

(2009).

47. Kosiuk, J. et al. Left ventricular diastolic dysfunction in atrial fibrillation: predictors and

relation with symptom severity. J. Cardiovasc. Electrophysiol. 23, 1073–1077 (2012).

48. Harrell, F. E., Jr, Lee, K. L. & Mark, D. B. Multivariable prognostic models: issues in

developing models, evaluating assumptions and adequacy, and measuring and reducing errors.

Stat. Med. 15, 361–387 (1996).

49. Wilson, P. W. et al. Prediction of coronary heart disease using risk factor categories.

Circulation 97, 1837–1847 (1998).

50. Conroy, R. M. et al. Estimation of ten-year risk of fatal cardiovascular disease in Europe: the

SCORE project. Eur. Heart J. 24, 987–1003 (2003).

51. Mancia, G. et al. 2013 ESH/ESC guidelines for the management of arterial hypertension: the

Task Force for the Management of Arterial Hypertension of the European Society of

Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur. Heart J. 34, 2159–

2219 (2013).

72

52. Dalen, H. et al. Segmental and global longitudinal strain and strain rate based on

echocardiography of 1266 healthy individuals: the HUNT study in Norway. Eur. J.

Echocardiogr. J. Work. Group Echocardiogr. Eur. Soc. Cardiol. 11, 176–183 (2010).

53. Chung, A. K. et al. Women have higher left ventricular ejection fractions than men independent

of differences in left ventricular volume: the Dallas Heart Study. Circulation 113, 1597–1604

(2006).

54. Biering-Sørensen, T. et al. Usefulness of the myocardial performance index determined by

tissue Doppler imaging m-mode for predicting mortality in the general population. Am. J.

Cardiol. 107, 478–483 (2011).

55. Weidemann, F. et al. Myocardial function defined by strain rate and strain during alterations in

inotropic states and heart rate. Am. J. Physiol. Heart Circ. Physiol. 283, H792–799 (2002).

56. Tei, C., Nishimura, R. A., Seward, J. B. & Tajik, A. J. Noninvasive Doppler-derived

myocardial performance index: correlation with simultaneous measurements of cardiac

catheterization measurements. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 10,

169–178 (1997).

57. Meric, M. et al. Tissue doppler myocardial performance index in patients with heart failure and

its relationship with haemodynamic parameters. Int. J. Cardiovasc. Imaging (2014).

doi:10.1007/s10554-014-0449-1

58. Lamb, H. J. et al. Diastolic dysfunction in hypertensive heart disease is associated with altered

myocardial metabolism. Circulation 99, 2261–2267 (1999).

59. Nakashima, Y., Nii, T., Ikeda, M. & Arakawa, K. Role of left ventricular regional

nonuniformity in hypertensive diastolic dysfunction. J. Am. Coll. Cardiol. 22, 790–795 (1993).

60. Cacciapuoti, F., Scognamiglio, A., Paoli, V. D., Romano, C. & Cacciapuoti, F. Left Atrial

Volume Index as Indicator of Left Ventricular Diastolic Dysfunction: Comparation between

73

Left Atrial Volume Index and Tissue Myocardial Performance Index. J. Cardiovasc.

Ultrasound 20, 25–29 (2012).

61. Akintunde, A. A., Akinwusi, P. O. & Opadijo, G. O. Relationship between Tei index of

myocardial performance and left ventricular geometry in Nigerians with systemic hypertension.

Cardiovasc. J. Afr. 22, 124–127 (2011).

62. Lewington, S. et al. Age-specific relevance of usual blood pressure to vascular mortality: a

meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 360,

1903–1913 (2002).

63. Levy, D., Garrison, R. J., Savage, D. D., Kannel, W. B. & Castelli, W. P. Prognostic

implications of echocardiographically determined left ventricular mass in the Framingham

Heart Study. N. Engl. J. Med. 322, 1561–1566 (1990).

64. Koren, M. J., Devereux, R. B., Casale, P. N., Savage, D. D. & Laragh, J. H. Relation of left

ventricular mass and geometry to morbidity and mortality in uncomplicated essential

hypertension. Ann. Intern. Med. 114, 345–352 (1991).

65. Vasan, R. S., Larson, M. G., Leip, E. P., Kannel, W. B. & Levy, D. Assessment of frequency of

progression to hypertension in non-hypertensive participants in the Framingham Heart Study: a

cohort study. Lancet 358, 1682–1686 (2001).

66. Vasan, R. S. et al. Impact of high-normal blood pressure on the risk of cardiovascular disease.

N. Engl. J. Med. 345, 1291–1297 (2001).

67. Møller, J. E. et al. Prognostic importance of systolic and diastolic function after acute

myocardial infarction. Am. Heart J. 145, 147–153 (2003).

68. Møller, J. E., Søndergaard, E., Poulsen, S. H. & Egstrup, K. The Doppler echocardiographic

myocardial performance index predicts left-ventricular dilation and cardiac death after

myocardial infarction. Cardiology 95, 105–111 (2001).

74

69. Arnlöv, J., Ingelsson, E., Risérus, U., Andrén, B. & Lind, L. Myocardial performance index, a

Doppler-derived index of global left ventricular function, predicts congestive heart failure in

elderly men. Eur. Heart J. 25, 2220–2225 (2004).

70. Arnlöv, J. et al. A Doppler-derived index of combined left ventricular systolic and diastolic

function is an independent predictor of cardiovascular mortality in elderly men. Am. Heart J.

149, 902–907 (2005).

71. Dujardin, K. S. et al. Prognostic value of a Doppler index combining systolic and diastolic

performance in idiopathic-dilated cardiomyopathy. Am. J. Cardiol. 82, 1071–1076 (1998).

72. Correale, M. et al. Tissue Doppler time intervals predict the occurrence of rehospitalization in

chronic heart failure: data from the daunia heart failure registry. Echocardiogr. Mt. Kisco N 29,

906–913 (2012).

73. Biering-Sørensen, T. et al. Prognostic Value of Cardiac Time Intervals by Tissue Doppler

Imaging M-Mode in Patients With Acute ST-Segment-Elevation Myocardial Infarction Treated

With Primary Percutaneous Coronary Intervention. Circ. Cardiovasc. Imaging 6, 457–465

(2013).

74. Biering-Sørensen, T. et al. Doppler Tissue Imaging Is an Independent Predictor of Outcome in

Patients with ST-Segment Elevation Myocardial Infarction Treated with Primary Percutaneous

Coronary Intervention. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 27, 258–

267 (2014).

75. Andersson, C. et al. Importance and inter-relationship of tissue Doppler variables for predicting

adverse outcomes in high-risk patients: an analysis of 388 diabetic patients referred for

coronary angiography. Eur. Heart J. Cardiovasc. Imaging 13, 643–649 (2012).

75

76. Andersson, C. et al. The prognostic importance of a history of hypertension in patients with

symptomatic heart failure is substantially worsened by a short mitral inflow deceleration time.

BMC Cardiovasc. Disord. 12, 30 (2012).

77. MacMahon, S. et al. Blood pressure, stroke, and coronary heart disease. Part 1, Prolonged

differences in blood pressure: prospective observational studies corrected for the regression

dilution bias. Lancet 335, 765–774 (1990).

78. Braunwald, E. & Kloner, R. A. The stunned myocardium: prolonged, postischemic ventricular

dysfunction. Circulation 66, 1146–1149 (1982).

79. Mogelvang, R. et al. Discriminating between cardiac and pulmonary dysfunction in the general

population with dyspnea by plasma pro-B-type natriuretic peptide. J. Am. Coll. Cardiol. 50,

1694–1701 (2007).

80. Su, H.-M. et al. Correlation of Tei index obtained from tissue Doppler echocardiography with

invasive measurements of left ventricular performance. Echocardiogr. Mt. Kisco N 24, 252–257

(2007).

Usefulness of the Myocardial Performance Index Determined byTissue Doppler Imaging M-Mode for Predicting Mortality in the

General Population

Tor Biering-Sørensen, MBa,b,*, Rasmus Mogelvang, MD, PhDa,c, Sune Pedersen, MD, PhDb,Peter Schnohr, MD, DMSca, Peter Sogaard, MD, DMScb, and Jan Skov Jensen, MD, DMSca,b

The objective of this study was to evaluate the prognostic value of the myocardial perfor-mance index (MPI), assessed by color-coded tissue Doppler imaging (TDI) M-modethrough the anterior mitral leaflet. Color TDI M-mode through the mitral leaflet is an easy,very fast, and precise method to estimate cardiac intervals and thus obtain the MPI, but thediagnostic and prognostic values of this parameter are unknown. In a large populationstudy, cardiac function was evaluated in 1,100 participants by conventional echocardiog-raphy and TDI. MPI was calculated from pulse-wave Doppler analyses of left ventricularin- and outflow using standard procedures (MPIconv) and by color-coded TDI M-modethrough the mitral leaflet in the apical 4-chamber view (MPITDI). MPITDI was increased insubjects with coronary heart disease (CHD) compared to controls, even after multivariableadjustment (p <0.002). During follow-up (median 5.3 years), 90 participants died. MPITDIwas significantly associated with overall mortality, and risk of dying increased by 31% per0.1 increase in MPITDI. In contrast to MPIconv, MPITDI provided independent prognosticinformation in a multivariable Cox proportional hazard model (adjusting for age, gender,body mass index, heart rate, mean arterial blood pressure, and CHD), with a hazard ratioof 1.18 (p � 0.01) per 0.1 increase in MPITDI. In conclusion, MPITDI is a quick, simple, andreproducible measurement, which is increased in subjects with CHD and provides inde-pendent prognostic information in a low-risk population. © 2011 Elsevier Inc. All rightsreserved. (Am J Cardiol 2011;107:478–483)

Cardiac intervals have been suggested and investigatedas potential markers of cardiac dysfunction for a longtime.1–4 Preservation of normal cardiac intervals is inti-mately related to normal cardiac physiology, hemodynam-ics, and mechanics. In the ailing myocardium cardiac inter-vals may change early in the disease process, thus providingearly signs of subclinical disease. The myocardial perfor-mance index (MPI),2,3 also known as the Tei index, is anattempt to expand the use of such intervals to characterizesystolic and diastolic performances by 1 single value de-rived from calculations of pulse-Doppler echocardiographyof the left ventricular (LV) outflow tract and mitral valveinflow. MPI has been shown to correlate significantly withestablished indexes of LV systolic and diastolic functionsby cardiac catheterization.5–7 Besides being a feasible non-invasive indicator of overall LV function, MPI has beenshown to provide prognostic value in the general population

and in patients with various cardiac conditions.8–12 As analternative noninvasive approach to obtain MPI, tissueDoppler imaging (TDI) of the mitral annular myocardialsegments (MPIregional) have been demonstrated to correlatewell with conventional MPI obtained by the pulse-Dopplermethod (MPIconv),13,14 although the absolute value obtainedby the 2 methods are different.15–17 TDI allows the intervalsto be obtained in the same cardiac cycle in contrast to theconventional method, which probably makes MPI as as-sessed by TDI less susceptible to variation in heart rate.MPIregional is, however, limited by regional myocardial dif-ferences in intervals. This could be overcome by analyzingglobal intervals through evaluation of mitral valve motionby a simple M-mode analysis18 instead of focusing onregional myocardial movement. Thus, using color TDIM-mode through the mitral valve to estimate MPI (MPITDI)may be a better method reflecting global cardiac time inter-vals and eliminating beat-to-beat variation.

Methods

This study was performed as a substudy of the FourthCopenhagen City Heart Study, a longitudinal cohort studyof cardiovascular disease and risk factors.19 The presentreport includes 1,100 randomly selected men and women(20 to 93 years old) who underwent an echocardiographicexamination including color TDI. Whether or not a partic-ipant underwent echocardiography was completely indepen-dent of his or her health status and other risk factors.Patients with atrial fibrillation, significant valvular stenosis,

aCopenhagen City Heart Study, Epidemiological Research Unit, Bispe-bjerg Hospital, Copenhagen, Denmark; bDepartment of Cardiology, Gen-tofte Hospital, Copenhagen, Denmark; cDepartment of Medicine, HolbaekHospital, Faculty of Health Sciences, University of Copenhagen, Copen-hagen, Denmark. Manuscript received May 31, 2010; revised manuscriptreceived and accepted September 22, 2010.

This study was financially supported by the Research Council forHealth and Disease, The Ministry of Science, Technology and Innovation,Copenhagen, Denmark.

*Corresponding author: Tel: 45-2-893-3590; fax: 45-3-977-7381.E-mail address: tor_biering@hotmail.com (T. Biering-Sørensen).

0002-9149/11/$ – see front matter © 2011 Elsevier Inc. All rights reserved. www.ajconline.orgdoi:10.1016/j.amjcard.2010.09.044

or regurgitation were excluded. All subjects gave writteninformed consent to participate, and the study was per-formed in accordance with the second Declaration ofHelsinki and approved by the regional ethics committee.

Coronary heart disease (CHD) was defined as a historyof hospital admission due to acute coronary artery occlu-sion, percutaneous coronary intervention, or coronaryartery bypass grafting or major ischemic alterations onelectrocardiogram as defined by Minnesota Codes 1.1 to3. Participants without CHD were defined as controls.

Three experienced echocardiographic technicians using aVingmed Ultrasound Vivid Five (GE Medical, Horten, Nor-way) with a 2.5-MHz probe recorded all echocardiogramsfor the Copenhagen City Heart Study. All subjects wereexamined in the left lateral decubitus position. All imageswere recorded using second harmonic imaging at the time ofend-expiration. Collected data were stored on magneto-optical disks and an external hard drive (FireWire, LaCie,France) and analyzed off-line with commercially availablesoftware (EchoPac, GE Medical), with the investigator be-ing blinded to other information.

Pulse-wave Doppler at the apical position was used torecord mitral inflow between the tips of the mitral leafletsand LV outflow just below the aortic annulus. One observerperformed conventional echocardiographic analyses. Theinterval from mitral valve closing to opening was deter-mined as the period from the end to onset of mitral inflow.LV ejection time (ET) was determined as the period fromonset to end of LV outflow. MPIconv was calculated as(interval from mitral valve closing to opening minus ejec-tion time)/ejection time.

Intervals by the TDI method were measured in the apical4-chamber view at the highest possible frame rate (median133 frames/s, interquartile range 37) using a 2 to 4-cmstraight M-mode line placed through the septal half of themitral leaflet, perpendicular to the leaflet in end-systole inclosed position18 (Figure 1).

One observer performed TDI analyses. From the colorM-mode diagram intervals were measured. TDI intervalfrom mitral valve closing to opening was defined as theperiod from color shift from blue to red at end-diastole tothe rapid opening of the mitral valve indicated by the colorshift of red-orange to yellow at the beginning of diastole(Figure 1). TDI ejection time was defined as the period fromthe color shift of blue to red color in systole ending at thecolor shift from red to blue at end-systole (Figure 1).MPITDI was calculated from measured cardiac intervals as(TDI interval from mitral valve closing to opening minusTDI ejection time)/TDI ejection time.

Participants were followed from the examination in 2002through 2003 until August 2007 or time of death using theunique personal identification number in the Central Officeof Civilian Registration. Follow-up was 100% complete.

Table 1 presents comparisons between groups performedby Student’s t test or Fisher’s exact test. Values in paren-theses are 95% confidence intervals unless otherwise stated.Intra- and interobserver variabilities were determined byBland-Altman plots of repeated analyses of stored echocar-diographic loops (n � 25). Association with MPIconv andMPITDI was tested for the defined groups by univariable andmultivariable regression analyses including the prespecifiedbaseline variables age, gender, body mass index, heart rate,and mean arterial blood pressure.

Tertiles of MPITDI were constructed and cumulative sur-vival curves were established by the Kaplan-Meier method.Univariable and multivariable Cox proportional hazards re-gression models were used to examine the association ofeach of the 2 methods used to obtain MPI and baselinevariables to risk of death. Mis-specification of the functionalform of covariates and the assumption of proportional haz-ards were evaluated by plots of cumulative martingale re-siduals.

A p value �5% in 2-sided tests was considered statisti-cally significant. All analyses were performed with SAS9.1.3 for Windows (SAS Institute, Cary, North Carolina).

Figure 1. (Left) Four-chamber gray-scale (bottom) and color tissue Dopplerimaging (top) views in end-systole showing the position of the M-modeline used for measurement of cardiac intervals. (Right) Color diagram ofthe tissue Doppler imaging M-mode of the anterior mitral leaflet. AVC �aortic valve closing; AVO � aortic valve opening; ET � ejection time;MVC � mitral valve closing; MVO � mitral valve opening; TST �interval from mitral valve closing to opening.

Table 1Characteristics of controls and participants with coronary heart disease

Variable Controls CHD p Value(n � 943) (n � 93)

Age (years) 59 � 16 69 � 12 �0.001Men 42% 37% 0.32Body mass index (kg/m2) 25.5 � 3.9 26.7 � 4.4 �0.01Heart rate (beats/min) 69 � 11 69 � 11 0.76Systolic blood pressure

(mm Hg)136 � 23 145 � 26 �0.001

Diastolic blood pressure(mm Hg)

78 � 12 79 � 15 0.47

Mean arterial bloodpressure (mm Hg)

97 � 14 101 � 16 �0.01

Conventional myocardialperformance index

0.39 � 0.14 0.45 � 0.20 �0.001

Tissue Doppler imagingmyocardialperformance index

0.51 � 0.12 0.59 � 0.22 �0.001

Dead at follow-up 7.5% 20.4% �0.001

Values are means � SDs or percentages of subjects.

479Methods/MPI by TDI M-Mode Predicts Death

Results

In total 1,100 subjects were examined with TDI andconventional echocardiography in this study. Forty-fivewere excluded due to inadequate quality of echocardio-graphic examination, and 19 were excluded due to atrialfibrillation or significant valvulopathy. Characteristics ofthe remaining subjects constituting the study population(n � 1,036) are listed in Table 1. The study population wasdivided into groups with and without CHD at time of en-rollment.

MPIconv remained significantly higher in the CHD groupafter multivariable adjustment for age, gender, body massindex, heart rate, and mean arterial blood pressure (0.44,0.41 to 0.48, vs 0.39, 0.38 to 0.40, p �0.002). This patternwas the same for MPITDI, which remained significantlyhigher in the CHD group after multivariable adjustment forthe same variables (0.56, 0.53 to 0.59, vs 0.51, 0.50 to 0.52,p �0.002). Besides being associated with the presence ofCHD, MPITDI increased independently with increasing ageand mean arterial blood pressure, whereas MPIconv in-creased independently with increasing age and mean arterialblood pressure and decreasing body mass index and heartrate (Table 2).

Risk of mortality increased with increasing tertiles ofMPITDI (Figure 2), being �3 times as high in the thirdtertile compared to the first tertile (hazard ratio 3.3, 1.8 to6.2, p �0.001).

MPI obtained from the 2 methods was significantly as-sociated with overall mortality (Table 3), and MPITDI re-mained independently associated with overall mortality af-ter multivariable adjustment for age, gender, body massindex, heart rate, mean arterial blood pressure, and CHD(Table 3).

Bland-Altman plots of intra- and interobserver differ-ences of MPITDI and MPIconv are shown in Figure 3. ForMPITDI intraobserver variability analysis showed a meandifference � SD of 0.016 � 0.038 and interobserver vari-ability analysis showed a mean difference of 0.010 � 0.073.For MPIconv intraobserver variability analysis showed amean difference of 0.013 � 0.057 and interobserver vari-ability analysis showed a mean difference of 0.082 � 0.114.

Discussion

In the present large prospective study we found thatMPIconv or MPITDI was significantly increased in subjects

with CHD compared to controls, even after multivariableadjustment. Furthermore, high MPIconv and MPITDI wereassociated with overall mortality and MPITDI providedprognostic information incremental to known confounderssuch as age, gender, body mass index, heart rate, meanarterial blood pressure, and CHD.

MPIconv is increased in subjects with a large variety ofcardiac diseases,3,11,15,16,18,20 including CHD.21–25 Ourstudy is the first to show that CHD is associated with a highMPITDI obtained with color TDI M-mode through the an-terior mitral valve leaflet. This result is interesting whenconsidering the possible advantages of MPITDI compared toMPIconv. Color TDI M-mode through the mitral leaflet is aneasy and very fast method to estimate MPITDI

18 and easy tolearn. Even when good imaging quality is difficult to obtain,it is often possible to visualize the mitral valve in the apicalview and assess its longitudinal movement by color TDIM-mode. Clear color shifts marking minimal changes indirection of the mitral valve are very easy to identify (Figure 1)and provides MPITDI with a potentially better reproducibil-ity than MPIconv (Figure 3). The latter is calculated fromintervals obtained from velocity curves where the curvesignals may be scattered, which can make it hard to accu-

Table 2Multiple regression models with conventional myocardial performance index and tissue Doppler imaging myocardial performance index as dependentvariables

MPIconv MPITDI

Beta (95% CI) p Value Beta (95% CI) p Value

Age per 10 years 0.010 (0.003 to 0.017) 0.006 0.025 (0.019 to 0.031) �0.0001Male gender 0.019 (�0.001 to 0.040) 0.06 0.004 (�0.013 to 0.021) 0.64Body mass index per 5 kg/m2 �0.015 (�0.028 to �0.001) 0.033 0.002 (�0.009 to 0.013) 0.77Heart rate per 10 beats/min �0.015 (�0.024 to �0.005) 0.002 0.006 (�0.001 to 0.014) 0.11Mean arterial blood pressure per 10 mm Hg 0.013 (0.005 to 0.021) 0.002 0.018 (0.011 to 0.025) �0.0001

Two multiple regression models with conventional myocardial performance index and tissue Doppler imaging myocardial performance index as dependentvariables and listed parameters as independent variables.

CI � confidence interval.

Figure 2. Kaplan-Meier survival curves depicting cumulative probability ofsurvival for subjects stratified into tertiles (1, �0.45; 2, �0.55 to �0.45; 3,�0.55) of myocardial performance index assessed by color tissue Dopplerimaging M-mode through the mitral leaflet.

480 The American Journal of Cardiology (www.ajconline.org)

rately define the cardiac intervals. Furthermore, MPITDI isobtained in 1 cardiac cycle and from 1 standard projection(apical 4-chamber view). MPIconv is obtained from �2

cardiac cycles from �1 projection, which makes MPIconvmore time consuming and prone to error caused by heartrate variability. We recognize that it is possible to obtain an

Table 3Conventional myocardial performance index and tissue Doppler imaging myocardial performance index as predictors of mortality by Cox proportionalhazards regression models

Univariable Multivariable Adjustment*

Hazard Ratio(95% CI)

p Value Hazard Ratio(95% CI)

p Value

Conventional myocardial performance index per 0.1 increase 1.20 (1.04–1.38) 0.010 1.11 (0.96–1.27) 0.15Tissue Doppler imaging myocardial performance index per

0.1 increase1.31 (1.19–1.44) �0.0001 1.18 (1.04–1.34) 0.010

* Adjustment for age, gender, body mass index, heart rate, mean arterial blood pressure, and ischemic heart disease.Abbreviation as in Table 2.

Figure 3. Bland-Altman plots of intra- and interobserver differences of myocardial performance index assessed by color tissue Doppler imaging M-mode throughthe mitral leaflet versus conventional myocardial performance index showing mean difference (solid lines) and 95% limits of agreement (dotted lines).

481Methods/MPI by TDI M-Mode Predicts Death

MPI from LV in- and outflow signals from 1 projection, i.e.,from 5-chamber view, but this method is not commonlyused in scientific evaluations and does not solve the diffi-culties by defining the cardiac intervals precisely.

Absolute values of MPIconv and MPITDI are different,which has also been found in previous studies comparingMPIconv to MPIregional.

15–17 MPITDI is based on analysis ofthe mitral leaflet moving passively depending on shifts inpressure and blood flow in the left ventricle and left atrium,whereas MPIconv is derived purely from blood flow. Itseems reasonable that the difference in physical parametersthat define the 2 methods may explain the discrepancybetween absolute values of MPIconv and MPITDI.

With progression of diastolic dysfunction, isovolumicrelaxation time prolongs as relaxation deteriorates and laterisovolumic relaxation time shortens when LV stiffness isincreased. This bidirectional course of isovolumic relaxationtime during development of diastolic dysfunction has beenmentioned as a major limitation in using MPIconv as a predictorof global cardiac function.23,25,26 Su et al27 investigated if theMPIregional could be used to evaluate degree of diastolicdysfunction. They found that conventional Doppler param-eters E, A, E/A ratio, and E deceleration time, as expected,displayed a bimodal distribution, whereas the TDI parame-ters e=, E/e=, and MPIregional showed unimodal distributionwith increasing severity of LV diastolic dysfunction.27 Thesame group also demonstrated that invasive measurementsfor systolic and diastolic functions, ejection fraction, andLV relaxation time constant (�) were independent predictorsof MPIregional.

5 These results provide support MPI obtainedby TDI as a feasible indicator in assessing ailing myocardialfunction.

Besides being associated with presence of CHD, ourresults showed that MPITDI was independently associatedwith age and mean arterial blood pressure, whereas MPIconvwas independently associated with age, body mass index,heart rate, and mean arterial blood pressure. These resultsare different from findings in previous studies where MPIwas shown to be independent of age, heart rate, and bloodpressure.3,5,23 Cui et al28 also demonstrated a correlationbetween MPIregional and age and heart rate, but althoughstatistically significant, the investigators found that changesof MPI were clinically insignificant. This also seems to besomewhat true for our findings (Table 2), where MPITDIwas found to be significantly associated with age and meanarterial blood pressure, but this association seems of lessclinical significance because beta coefficients were found tobe 0.025 (0.019 to 0.031) and 0.018 (0.011 to 0.025) per 10change in age or blood pressure, respectively, in our mul-tiple regression model. Although we found MPITDI to bedependent on age and mean arterial blood pressure, MPITDIhad predictive value independent of these factors.

MPIconv was influenced by several physical parameterscompared to MPITDI; hence, it seems as if MPI based onsimple cardiac mechanics are less confounded than MPIbased on calculations from versatile blood flow patternsaround the mitral valve, which may make the interpretationof MPITDI less complex.

MPIconv has been demonstrated to be an independentpredictor of cardiovascular mortality in a population ofelderly men after adjusting for LV hypertrophy, CHD, and

other traditional cardiovascular risk factors.9 Furthermore,MPIconv appeared to be a superior predictor of cardiovas-cular mortality and congestive heart failure compared toother echocardiographic measurements of cardiac function(ejection fraction, wall motion index, and E/A ratio) andcardiovascular risk factors.8,9 Superiority of the predictivecapacity of MPIconv has also been demonstrated in a largepopulation after acute myocardial infarction.24 In this studyMoller et al24 showed that MPIconv and a restrictive fillingpattern were independent predictors of all-cause death,whereas wall motion index and a filling pattern of impairedrelaxation were statistically insignificant. These results em-phasize the prognostic information we can gain by combin-ing information on systolic and diastolic performances. Theeas-index,19 like the MPITDI, is intended to incorporate theinterdependent and coherent relations between systolic anddiastolic functions in a simple feasible index that can pro-vide significant diagnostic and prognostic information.However, although the eas-index was significantly associ-ated with overall mortality in the general population andsuperior to other TDI and conventional echocardiographic pa-rameters,19 it is much more time consuming and difficult toobtain because myocardial velocity curves from 6 mitral an-nular segments obtained from all 3 apical projections are re-quired. This makes MPITDI more feasible in clinical practice.

Our study is the first to evaluate the prognostic value ofMPITDI in the general population. Previous studies, whichevaluated the prognostic value of MPIconv, were performedin selected populations, e.g., patients after acute myocardialinfarction,22,24 elderly men,8,9 American Indians,29 and pa-tients with cardiac amyloidosis,11 whereas our study popu-lation was randomly selected from the general population.

Our results show that MPITDI is superior to MPIconvbecause only MPITDI predicts overall mortality in a low-riskgeneral population after adjusting for age, gender, bodymass index, heart rate, mean arterial blood pressure, andCHD. Kaplan-Meier curves depicting the cumulative prob-ability of survival for subjects stratified into tertiles ofMPITDI (Figure 2) provide a visual impression of the sig-nificant impact of MPITDI on overall mortality. There is adifference of only 0.1 in MPITDI between participants in thelowest tertile (�0.45) and those in the highest tertile(�0.55), but the survival percentages are markedly differ-ent. These results seem promising when considering ourfindings with high reproducibility of MPITDI (Figure 3).Previous validation of the method showed that the intervalsobtained with the color TDI M-mode line through the an-terior mitral leaflet was an accurate and reproduciblemethod, especially when intervals were used to calculateMPITDI.

18 In concordance with these findings, we found thatthe precision of the measurement of MPI improved whenmeasured with the M-mode TDI method.

Based on our results MPITDI cannot be interpreted as anall explanatory echocardiographic parameter but rather as aquick, easy, and reproducible predictor that can be used forfast risk stratification. MPITDI can add to the informationacquired from other echocardiographic parameters, whichoften may be ambiguous especially in a low-risk population.

MPIconv has also been found to be significantly associ-ated with indexes of peripheral resistance and arterial com-pliance.29,30 This indicates that, in addition to reflecting LV

482 The American Journal of Cardiology (www.ajconline.org)

function, MPI might reflect peripheral pathology. This ar-terial-ventricular coupling reflected in the MPI adds a wholenew interesting dimension to analyses of cardiac intervals.

Certain limitations have to be taken into account for thepresent study. First, the inhabitants of Denmark and the studypopulation are primarily Caucasian, which limits the general-izability of our findings to other general populations with othercomposition. Our results were obtained in a low-risk popu-lation and may not be extrapolated to high-risk populationsin hospital settings. Second, we did not investigate cause ofdeath or why we could not discriminate between cardiovas-cular causes and other causes of death. However, we assumethat subjects with cardiac dysfunction found by echocardi-ography are more likely to die due to cardiovascular causesand that, if we were able to limit our analysis to cardiovas-cular deaths, the prognostic impact of MPITDI would beeven higher.

1. Mancini GB, Costello D, Bhargava V, Lew W, LeWinter M, KarlinerJS. The isovolumic index: a new noninvasive approach to the assess-ment of left ventricular function in man. Am J Cardiol 1982;50:1401–1408.

2. Tei C. New non-invasive index for combined systolic and diastolicventricular function. J Cardiol 1995;26:135–136.

3. Tei C, Ling LH, Hodge DO, Bailey KR, Oh JK, Rodeheffer RJ, TajikAJ, Seward JB. New index of combined systolic and diastolic myo-cardial performance: a simple and reproducible measure of cardiacfunction—a study in normals and dilated cardiomyopathy. J Cardiol1995;26:357–366.

4. Weissler AM, Harris WS, Schoenfeld CD. Systolic time intervals inheart failure in man. Circulation 1968;37:149–159.

5. Su HM, Lin TH, Voon WC, Lee KT, Chu CS, Yen HW, Lai WT, SheuSH. Correlation of Tei index obtained from tissue Doppler echocar-diography with invasive measurements of left ventricular performance.Echocardiography 2007;24:252–257.

6. Tei C, Nishimura RA, Seward JB, Tajik AJ. Noninvasive Doppler-derived myocardial performance index: correlation with simultaneousmeasurements of cardiac catheterization measurements. J Am SocEchocardiogr 1997;10:169–178.

7. LaCorte JC, Cabreriza SE, Rabkin DG, Printz BF, Coku L, WeinbergA, Gersony WM, Spotnitz HM. Correlation of the Tei index withinvasive measurements of ventricular function in a porcine model.J Am Soc Echocardiogr 2003;16:442–447.

8. Arnlov J, Ingelsson E, Riserus U, Andren B, Lind L. Myocardialperformance index, a Doppler-derived index of global left ventricularfunction, predicts congestive heart failure in elderly men. Eur Heart J2004;25:2220–2225.

9. Arnlov J, Lind L, Andren B, Riserus U, Berglund L, Lithell H. ADoppler-derived index of combined left ventricular systolic and dia-stolic function is an independent predictor of cardiovascular mortalityin elderly men. Am Heart J 2005;149:902–907.

10. Dujardin KS, Tei C, Yeo TC, Hodge DO, Rossi A, Seward JB.Prognostic value of a Doppler index combining systolic and diastolicperformance in idiopathic-dilated cardiomyopathy. Am J Cardiol1998;82:1071–1076.

11. Tei C, Dujardin KS, Hodge DO, Kyle RA, Tajik AJ, Seward JB.Doppler index combining systolic and diastolic myocardial perfor-mance: clinical value in cardiac amyloidosis. J Am Coll Cardiol1996;28:658–664.

12. Moller JE, Sondergaard E, Poulsen SH, Egstrup K. The Dopplerechocardiographic myocardial performance index predicts left-ventric-ular dilation and cardiac death after myocardial infarction. Cardiology2001;95:105–111.

13. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Value ofmeasuring myocardial performance index by tissue Doppler echocar-diography in normal and diseased heart. Jpn Heart J 2003;44:403–416.

14. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Novel approachto measure myocardial performance index: pulsed-wave tissue Dopp-ler echocardiography. Echocardiography 2003;20:503–510.

15. Gaibazzi N, Petrucci N, Ziacchi V. Left ventricle myocardial perfor-mance index derived either by conventional method or mitral annulustissue-Doppler: a comparison study in healthy subjects and subjectswith heart failure. J Am Soc Echocardiogr 2005;18:1270–1276.

16. Duzenli MA, Ozdemir K, Aygul N, Soylu A, Aygul MU, Gok H.Comparison of myocardial performance index obtained either by con-ventional echocardiography or tissue Doppler echocardiography inhealthy subjects and patients with heart failure. Heart Vessels 2009;24:8–15.

17. Harada K, Tamura M, Toyono M, Oyama K, Takada G. Assessment ofglobal left ventricular function by tissue Doppler imaging. Am JCardiol 2001;88:927–932.

18. Kjaergaard J, Hassager C, Oh JK, Kristensen JH, Berning J, SogaardP. Measurement of cardiac time intervals by Doppler tissue M-modeimaging of the anterior mitral leaflet. J Am Soc Echocardiogr 2005;18:1058–1065.

19. Mogelvang R, Sogaard P, Pedersen SA, Olsen NT, Marott JL, SchnohrP, Goetze JP, Jensen JS. Cardiac dysfunction assessed by echocardio-graphic tissue Doppler imaging is an independent predictor of mortal-ity in the general population. Circulation 2009;119:2679–2685.

20. Bruch C, Schmermund A, Marin D, Katz M, Bartel T, Schaar J, ErbelR. Tei-index in patients with mild-to-moderate congestive heart fail-ure. Eur Heart J 2000;21:1888–1895.

21. Dagdelen S, Eren N, Karabulut H, Caglar N. Importance of the indexof myocardial performance in evaluation of left ventricular function.Echocardiography 2002;19:273–278.

22. Poulsen SH, Jensen SE, Nielsen JC, Moller JE, Egstrup K. Serialchanges and prognostic implications of a Doppler-derived index ofcombined left ventricular systolic and diastolic myocardial perfor-mance in acute myocardial infarction. Am J Cardiol 2000;85:19–25.

23. Bruch C, Schmermund A, Dagres N, Katz M, Bartel T, Erbel R.Tei-Index in coronary artery disease—validation in patients with over-all cardiac and isolated diastolic dysfunction. Z Kardiol 2002;91:472–480.

24. Moller JE, Egstrup K, Kober L, Poulsen SH, Nyvad O, Torp-PedersenC. Prognostic importance of systolic and diastolic function after acutemyocardial infarction. Am Heart J 2003;145:147–153.

25. Nearchou NS, Tsakiris AK, Tsitsirikos MD, Karatzis EN, Lolaka MD,Flessa KD, Bogiatzis DT, Skoufas PD. Tei index as a method ofevaluating left ventricular diastolic dysfunction in acute myocardialinfarction. Hellenic J Cardiol 2005;46:35–42.

26. Gillebert TC, Van dV, De Buyzere ML, De SJ. Time intervals andglobal cardiac function. Use and limitations. Eur Heart J 2004;25:2185–2186.

27. Su HM, Lin TH, Voon WC, Lee KT, Chu CS, Lai WT, Sheu SH.Differentiation of left ventricular diastolic dysfunction, identificationof pseudonormal/restrictive mitral inflow pattern and determination ofleft ventricular filling pressure by Tei index obtained from tissueDoppler echocardiography. Echocardiography 2006;23:287–294.

28. Cui W, Roberson DA, Chen Z, Madronero LF, Cuneo BF. Systolic anddiastolic time intervals measured from Doppler tissue imaging: normalvalues and Z-score tables, and effects of age, heart rate, and bodysurface area. J Am Soc Echocardiogr 2008;21:361–370.

29. Mishra RK, Kizer JR, Palmieri V, Roman MJ, Galloway JM, FabsitzRR, Lee ET, Best LG, Devereux RB. Utility of the myocardial per-formance index in a population with high prevalences of obesity,diabetes, and hypertension: the strong heart study. Echocardiography2007;24:340–347.

30. Lind L, Andren B, Arnlov J. The Doppler-derived myocardial perfor-mance index is determined by both left ventricular systolic and dia-stolic function as well as by afterload and left ventricular mass.Echocardiography 2005;22:211–216.

483Methods/MPI by TDI M-Mode Predicts Death

1

Cardiac Time Intervals by Tissue Doppler

Imaging M-mode: Normal Values and association

with established Echocardiographic and Invasive

measures of Systolic and Diastolic function

Tor Biering-Sørensen*†‡, MD; Rasmus Mogelvang*†, MD, PhD;

Martina Chantal de Knegt*, MD; Flemming Javier Olsen*, MB; Søren Galatius*, MD, DMSc,

and Jan Skov Jensen*†‡, MD, PhD, DMSc.

*Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark

†The Copenhagen City Heart Study, Bispebjerg Hospital, University of Copenhagen, Denmark

‡Institute of Clinical Medicine, Faculty of Health Sciences, University of Copenhagen, Denmark

Running title:.Normal values of cardiac time intervals by TDI M-mode

Word count: 5,598 (includes title page, abstract, text, references, tables, and figure legends)

Correspondence:

Tor Biering-Sørensen, MD

Department of Cardiology, Gentofte University Hospital, Copenhagen, Denmark

Niels Andersensvej 65, DK-2900, Post 835, Copenhagen (Denmark),

Phone: +45 28933590, Fax: +45 39777381

E-mail: tor.biering@gmail.com

2

Abstract

Purpose: To define normal values of the cardiac time intervals obtained by tissue Doppler imaging

(TDI) M-mode through the mitral valve (MV). Furthermore, to evaluate the association of the

myocardial performance index (MPI) obtained by TDI M-mode (MPITDI) and the conventional

method of obtaining MPI (MPIConv), with established echocardiographic and invasive measures of

systolic and diastolic function.

Methods: In a large community based population study (n=989), where all are free of any

cardiovascular disease and cardiovascular risk factors, cardiac time intervals, including isovolumic

relaxation time (IVRT), isovolumic contraction time (IVCT), and ejection time (ET) were obtained

by TDI M-mode through the MV. IVCT/ET, IVRT/ET and the MPI (IVRT+IVCT)/ET) were

calculated. We also included a validation population (n=44) of patients who underwent left heart

catheterization and had the MPITDI and MPIConv measured.

Results: IVRT, IVRT/ET and MPI all increased significantly with increasing age in both genders

(p<0.001 for all). IVCT, ET, IVRT/ET, and MPI differed significantly between males and females,

displaying that women, in general exhibit better cardiac function.

MPITDI was significantly associated with invasive (dP/dt max) and echocardiographic measures of

systolic (LVEF, global longitudinal strain and global strainrate s) and diastolic function (e’, global

strainrate e)(p<0.05 for all), whereas MPIConv was significantly associated with LVEF, e’ and global

strainrate e (p<0.05 for all).

Conclusion: Normal values of cardiac time intervals differed between genders and deteriorated

with increasing age. The MPITDI (but not MPIConv) is associated with most invasive and established

echocardiographic measures of systolic and diastolic function.

3

Introduction

Cardiac time intervals are sensitive markers of cardiac dysfunction, even when it goes unrecognized

by conventional echocardiography(1–3). These subtle impairments in cardiac function, which are

revealed by disturbed cardiac time intervals, have previously been demonstrated to contain

prognostic information incremental to the conventional echocardiographic parameters(4–8). The

great potential of the cardiac time intervals to capture signs of cardiac dysfunction was already

demonstrated half a century ago(9). Since then, cardiac ultrasound machines have evolved and new

methods of obtaining cardiac time intervals have been suggested. The most common method of

obtaining cardiac time intervals by echocardiography is from velocity curves obtained from pulsed-

Doppler echocardiography of the left ventricular (LV) outflow tract and mitral valve (MV)

inflow(10, 11). It is hereby possible to calculate the index of combined systolic and diastolic

performance and is known as the myocardial performance index (MPIConv)(10, 11). Additionally,

using the newer echocardiographic modality, tissue Doppler imaging (TDI) which provides regional

myocardial velocity curves and, thereby, regional cardiac time intervals(12) and MPI (MPIRegional),

has in recent years been an attractive method in the obtainment of cardiac time intervals(13–15).

The two aforementioned methods of obtaining cardiac time intervals have, however, several

limitations which have been a limiting factor in the clinical applicability of these methods until

now. A novel method has recently evolved where global cardiac time intervals are obtained through

evaluation of MV movement through the cardiac cycle by using a simple color TDI M-mode

analysis(7, 16–18). Using color TDI M-mode through the MV to estimate cardiac time intervals is

an improved method, with better precision and reproducibility(17, 18), reflecting global cardiac

time intervals and eliminating beat-to-beat variation and regional differences.

Nevertheless, clear definitions of normal and pathological cardiac time intervals, including their

normal values, have to be established before we can use these in a clinical setting in order to be able

4

to identify miniscule levels of impaired cardiac function. In addition, the association between this

novel method of obtaining MPI (MPITDI) and established echocardiographic and invasive measures

of systolic and diastolic function has to be evaluated. The aim of this study was, therefore, to define

normal values of cardiac time intervals obtained by TDI M-mode through the MV in a large sample

obtained from the general population without known cardiovascular disease and cardiovascular risk

factors. A further aim of this study was to evaluate the association of MPITDI (the novel method of

obtaining MPI) and MPIconv (the conventional method of obtaining MPI) as described by Tei and

colleagues(10, 11) with established echocardiographic and invasive measures of systolic and

diastolic function.

5

Method

Study population: Our study population consists of individuals without hypertension, diabetes,

ischemic heart disease (IHD), heart failure (CHF) or atrial fibrillation who are included in the

echocardiographic sub study of the fourth Copenhagen City Heart Study (Study group A)(18–20),

and an invasive validation group consisting of patients suspected of ischemic heart disease who

underwent coronary angiography, left heart catheterization and echocardiography (Study group B).

Population contributing to the normal values (group A): This population includes all participants

from the fourth Copenhagen City Heart Study examination 2001-2003, who had an

echocardiographic examination, including TDI, performed and who were free from any

cardiovascular disease and cardiovascular risk factor (n = 989)(Table 1 and 2).

Invasive validation population (group B): The validation population includes patients suspected of

ischemic heart disease who underwent coronary angiography and left heart catheterization (n = 44)

at Department of Cardiology, Gentofte Hospital, University of Copenhagen between April 2009 and

February 2011.

Invasive Measurements: In group B, left heart catheterization was performed. Left ventricular

(LV) pressures were measured, and an echocardiogram was performed according to a standardized

protocol. Invasively measured pressure curves were obtained using 5 French pigtail catheters placed

in the LV chamber and the pressure set and transducer were calibrated. Measurements were

obtained over at least three cardiac cycles and saved as hard copies and digitalized in relation to this

study. The LV pressure curves were digitalized using Dagra Version 2.0.12.35924. In all cardiac

cycles the rate of LV pressure rise in early systole (dP/dt max) and the rate of LV pressure decline

in early diastole (dP/dt min) were measured and averaged over all the cardiac cycles.

6

Echocardiography: All echocardiograms were obtained using Vivid ultrasound systems (GE

Healthcare, Horten, Norway) with a 2.5-MHz transducer. All participants were examined with

conventional two-dimensional echocardiography and color TDI. All echocardiograms were stored

and analyzed offline with commercially available software (EchoPac, GE Medical, Horten,

Norway).

Conventional echocardiography: In group A (18–20), regional function was evaluated by the

standard 16-segment model, as suggested by the American Society of Echocardiography(21). LV

ejection fraction (LVEF) was evaluated by 1 observer on the basis of the wall motion index score.

In group B, LVEF was obtained using modified biplane Simpson’s method(21).

The myocardial thickness and the dimensions of the LV in end-diastole (LVIDd) and the left atrium

in end-systole (LAD) were measured. LV mass index (LVMI) was calculated as the anatomic

mass(21) divided by body surface area(22). Pulsed wave Doppler at the apical view was utilized to

record mitral inflow between the tips of the MVs. Peak velocity of early (E) and atrial (A) diastolic

filling and deceleration time of the E-wave (DT) were measured and the E/A-ratio was calculated.

In group B, pulsed-wave TDI tracings were obtained with the range gate placed at the septal and

lateral mitral annular segments in the 4-chamber view. The peak longitudinal early diastolic (e’)

velocity was measured and the average was calculated from the lateral and septal velocities and

used to obtain the E/e’. Impaired systolic function was defined as an LVEF below 50% and

impaired diastolic function as e’ below 9 cm/sec and/or an left atrium volume index of 34 mL/m2

or

above.

Additionally, in group B time intervals were measured using the method described by Tei and

colleagues(10, 11). The interval from MV closing to opening was determined as the period from the

end to onset of mitral inflow (A to E time) obtained from the pulsed-wave Doppler at the apical

7

position. LV ejection time (ET) was determined as the period from onset to end of LV outflow.

MPIConv was calculated as: (A to E time – ET)/ET(10, 11).

Two-dimensional strain echocardiography: In group B, Two-dimensional strain echocardiography

(2DSE) was performed from the apical 4-chamber, 2-chamber and apical long-axis view (mean 79

frames/s, standard deviation 18 frames/s). By speckle tracking, the endocardial border was traced in

end systole. The integrity of speckle tracking was automatically detected and visually ascertained.

In case of poor tracking, the region of interest tracing was readjusted. Global longitudinal strain

(GLS) was measured in all three apical views and averaged to provide GLS. Furthermore, in each

three apical views, global longitudinal systolic strain rate (GL SR s) and global longitudinal early

diastolic strain rate (GL SR e) were measured, and averaged to provide global estimates.

Cardiac Time Intervals obtained by color TDI M-mode: In group A (18–20), color TDI tracings

were obtained with the range gate placed at the lateral, septal, inferior, anterior, posterior and

anterior septal mitral annular segments in the 4-, 2- and 3-chamber view. e’ was measured by color

TDI and the average was calculated from the lateral, septal, inferior, anterior, posterior and anterior

septal velocities and used to obtain E/e’.

In both group A (18–20) and group B, cardiac time intervals were obtained by placing a 2-4 cm

straight M-mode line through the septal half of the MV in the color TDI 4-chamber view and

cardiac time intervals were measured directly from the color diagram(7, 18) (Figure 1). The IVCT

was defined as the time interval from MV closure (MVC), determined by the color shift from

blue/turquoise to red at end-diastole, to aortic valve opening (AVO) determined by the color shift

from blue to red (Figure 1). ET was defined as the time interval from AVO to aortic valve closing

(AVC), determined by the color shift from red to blue at end systole (Figure 1). IVRT was defined

8

as the time interval from AVC to MV opening (MVO), determined by the color shift from red-

orange to yellow (Figure 1). The method has previously been validated (7, 17, 18), and has been

demonstrated to be superior to the conventional method of obtaining cardiac time intervals(17, 18).

Furthermore, we have previously demonstrated high reproducibility of the method in this present

study population(18). Both isovolumic time intervals were divided with ET creating IVRT/ET and

IVCT/ET, respectively, and MPI was calculated as the sum of the two ((IVRT+IVCT)/ET).

Statistics: Proportions were compared using χ2-test, continuous Gaussian distributed variables with

Student’s t-test and Mann-Whitney test if non-Gaussian distributed. The association between

cardiac time intervals, age and gender was tested by univariable and multivariable regression

analyses including significant confounders detected from the baseline differences between genders

(Table 1). Linearity, variance homogeneity, and the assumption of normality were tested with plots

of residuals. Trends were analyzed by linear regression analyses and departure from linearity was

assessed by simultaneous assessment of linear and quadratic effects. Non-Gaussian distributed

continues variables (LVMI and E/e’) were categorised into quartiles when included in the models.

A p-value ≤ 0.05 in 2-sided test was considered statistically significant. All analyses were

performed with STATA Statistics/Data analysis, SE 12.0 (StataCorp, Texas,USA).

9

Results

Normal values of cardiac time intervals in healthy participants stratified by age and gender:

In participants without hypertension, diabetes, IHD, CHF or atrial fibrillation, normal values of

cardiac time intervals were assessed. The population that contributed to the normal values consisted

of 989 participants, 554 women and 427 men, (Table 1). The study participants were stratified

according to age (20 to 39, 40 to 59, and ≥ 60 years) and gender (Table 2 and Figure 2).

Gender significantly modified the relationship between age and two of the cardiac time intervals

(IVCT and IVCT/ET) which is visualized in Figure 2 (p for interaction = 0.004 for IVCT; p for

interaction = 0.005 for IVCT/ET). IVRT, IVRT/ET and MPI all increased significantly with

increasing age in both genders (p<0.001 for all)(Table 2 and Figure 2). IVCT and IVCT/ET

increased statistically significantly with increasing age in females but not in males (Table 2 and

Figure 2). ET decreased with increasing age in both genders; however, this decline was not

statistically significant (Table 2 and Figure 2). IVCT, ET, IVRT/ET, and MPI differed significantly

between male and female gender. These gender differences remained statistically significant after

multivariable adjustment for all other baseline differences between genders reported in Table 1.

The association between MPITDI and MPIConv, and invasive measures of systolic and diastolic

function: There was no association between MPIConv and dP/dt max (p=0.18; Table 3 and Figure

3a). The MPITDI decreased significantly with increasing values of dP/dt max (p=0.047; Table 3 and

Figure 3a), reflecting improved values of the combined cardiac index with improved cardiac

contractility. In comparison, there was no association between dP/dt max and GLS (p=0.20) or

LVEF (p=0.11). The only systolic echocardiographic parameter which displayed improved systolic

function with better LV contractility determined by higher dP/dt max was GL Strainrate s (GL

Strainrate s: β = -0.021 (-0.039 to -0.003), R2 = 0.12, p = 0.025, per 100 mmHg/sec increase).

10

Neither MPIConv nor MPITDI were significantly associated with dP/dt min (Table 3 and Figure 3b).

With regards to MPITDI we did, however, find a trend where MPITDI seemed to increase with

increasing values of dP/dt min (lower numerical values)(Table 3 and Figure 3a), reflecting a

worsening diastolic function (Table 3 and Figure 3b). In contrast, for MPIConv, if a trend was

present, it seemed reversed, as MPIConv increased with decreasing values of dP/dt min (higher

numerical values), reflecting improved diastolic function (Table 3 and Figure 3b). In comparison,

neither e’, GL Strainrate e, nor E/e’ were significantly associated with changes in dP/dt min (e’: β =

0.155 (-0.086 to 0.396), R2 = 0.04, p = 0.20, per 100 mmHg/sec decrease; GL Strainrate e: β =

0.005 (-0.021 to 0.031), R2 = 0.00, p = 0.71, per 100 mmHg/sec decrease; E/e’: β = -0.006 (-0.532

to 0.520), R2 = 0.00, p = 0.98, per 100 mmHg/sec decrease).

The association between MPITDI and MPIConv and established echocardiographic measures of

systolic and diastolic function: MPITDI increased significantly, displaying worsening myocardial

function with worsening systolic function determined by LVEF, GLS or GL Strainrate s (Table 3).

In contrast, MPIConv only increased significantly with worsening LVEF (Table 3).

Both MPITDI and MPIConv increased significantly with worsening diastolic function determined by e’

or GL Strainrate e (Table 3). Neither MPIConv nor MPITDI were significantly associated with E/e’

(Table 3).

Additionally, MPITDI was defined as increased when above 0.54 for women and 0.56 for men as

obtained from group A (Table 2). Furthermore, in group B an impaired systolic function was

defined as an LVEF below 50% and impaired diastolic function as e’ below 9 cm/sec (average

obtained from the septal and lateral wall by pulsed-wave TDI) and/or an left atrium volume index of

34 mL/m2

or above. Applying the cut-offs obtained from group A on group B, resulted in a

11

sensitivity of 26% and a specificity of 100% for the MPITDI to identify an impaired systolic and/or

diastolic function defined by conventional echocardiography (Table 4).

12

Discussion

Normal values of Cardiac time intervals stratified by age and gender: Our study is the first to

assess the normal values of cardiac time intervals obtained by TDI M-mode in a sample from the

general population without cardiovascular disease and cardiovascular risk factors (Table 2). We

found that LV function assessed by IVRT, IVRT/ET, and MPI, all decreased with increasing age in

both genders (Table 2 and Figure 2). The decrease in LV cardiac function with increasing age as

assessed by IVCT and IVCT/ET was only statistically significant in females and not in males

(Table 2 and Figure 2). ET decreased with increasing age in both genders; however, this decline

was not statistically significant (Table 2 and Figure 2). Other novel echocardiographic parameters

detecting subtle myocardial impairments like two-dimensional strain echocardiography(23) and

TDI(20) have previously, and in accordance with our results, been demonstrated to detect miniscule

LV dysfunction with increasing age despite the absence of any cardiovascular disease or risk

factors.

IVCT, ET, IVRT/ET, and MPI all differed significantly between male and female gender (Table 2

and Figure 2). These gender differences remained statistically significant after multivariable

adjustment for all other baseline differences between genders (Table 2). Gender differences exist

both in novel echocardiographic parameters like two-dimensional strain echocardiography(23) and

TDI(20), but also in the cardiac dimensions obtained by conventional echocardiography(21), and in

conventional measures of systolic(24) and diastolic function(21, 20). The discrepancy we found, i.e.

that women, in general, display better cardiac function than men, appears to be physiologically

plausible and in accordance with previous studies(21, 23, 20, 24).

The association between MPITDI and MPIConv and invasive and echocardiographic measures

of systolic and diastolic function: MPITDI has previously been demonstrated to have better

13

reproducibility(17, 18) and to be influenced by fewer physical parameters compared to MPIConv(18).

Additionally, MPITDI (but not MPIConv) has been proven to be an independent predictor of mortality

in the general population(18). This study is, however, the first to compare the two methods of

obtaining MPI with established invasive and echocardiographic measures of systolic and diastolic

performance. Our results demonstrate that MPITDI (but not MPIConv) is significantly correlated with

the invasive measure of contractility and systolic function dP/dt max (Table 3 and Figure 3a). The

only other echocardiographic measure which was significantly correlated with dP/dt max was GL

SR s, which previously has been demonstrated to be the most accurate marker of myocardial

contractile function(25). MPITDI may, therefore, be an accurate marker of myocardial contractile

function as well. In contrast to our results, Tei and colleagues previously demonstrated good

correlation between MPIConv and dP/dt max in a smaller study of patients with IHD(26). However,

in accordance with our results, several studies performed during the last decade which compare

MPI obtained by the conventional method as described by Tei and colleagues(10, 11) and by TDI,

have demonstrated MPI obtained by TDI to be the superior method(17, 18, 27, 28), both when

comparing reproducibility(17, 18) and their diagnostic and prognostic utilities(18, 27, 28).

Neither MPIConv nor MPITDI were significantly associated with dP/dt min in our study. In

comparison, none of the diastolic echocardiographic measures (e’, GL Strainrate e, nor E/e’) were

significantly associated with changes in dP/dt min. The lack in correlation between our invasive

measure of diastolic function (dP/dt min) and the echocardiographic measures of diastolic function

may, therefore, display an inherent limitation in the invasive dP/dt min parameter than in the

echocardiographic measures investigated in this study.

Furthermore, MPITDI increased significantly with worsening systolic function determined by all the

echocardiographic measures of systolic function (LVEF, GLS or GL SR s), whereas MPIConv only

increased significantly with worsening LVEF (Table 3). However, both MPITDI and MPIConv

14

increased with worsening diastolic function determined by the measures of diastolic function (e’ or

GL SR e) (Table 3). Neither MPITDI nor MPIConv correlated significantly with E/e’ (Table 3). This

may be due to the fact that E/e’ is an echocardiographic measure which correlates well with the LV

end diastolic pressure (LVEDP), but is not a sensitive measure of an ailing diastolic function when

LVEDP is normal. Likewise, the current guidelines recommend the evaluation of the value e’ (not

E/e’) as the first step in the assessment of grading diastolic dysfunction(29). Additionally, we have

previously demonstrated MPITDI to be independently associated with the conventional

echocardiographic measures of systolic and diastolic function (LVEF and e')(7). In addition, in the

present study we demonstrate that MPITDI, but not MPIConv, displays an ailing systolic and diastolic

function, as determined by the comparison with 2D-speckle tracking echocardiographic measures

(GLS, GL SR s and GL SR e). Our and previous results illustrate that an ailing systolic or diastolic

performance can be detected by an increasing value of MPI when assessed by TDI. We found when

the MPITDI was increased in our validation population (group B) above the reference limit obtained

from the healthy general population (group A), the probability of identifying an impaired

echocardiographic examination, defined by either an impaired systolic and/or diastolic function,

was 100%, illustrated by a positive predictive value of 100% (Table 4). Additionally, no one with a

normal echocardiography in our validation population had an MPI TDI above our defined reference

value illustrated by specificity of 100% (Table 4). Furthermore, MPITDI seem superior to MPIConv in

the detection of an ailing systolic or diastolic performance regardless of evaluation with

conventional echocardiographic, novel echocardiographic or invasive measures.

Limitations

15

The inhabitants of Denmark and the study population contributing to the normal values are

primarily Caucasian, which limits the generalizability of our findings to other general populations

with another race composition.

Our invasive validation population was small, comprised only 44 individuals, which limits our

study. Nevertheless, this was a larger study group than can be found in previous studies which

evaluate the association of cardiac time intervals with invasive measures of systolic and diastolic

function(26, 30).

Conclusion

The normal values of cardiac time intervals deteriorated with increasing age and differed between

genders, displaying that woman, in general, exhibit better cardiac function. MPITDI (but not

MPIConv) is significantly associated with most invasive and established echocardiographic measures

of systolic and diastolic function.

Funding Sources

This study was financially supported by the Faculty of Health Sciences, University of Copenhagen,

Denmark. The sponsor had no role in the study design, data collection, data analysis, data

interpretation, or writing of the manuscript.

16

References

1. Ruggiero A, De Rosa G, Rizzo D, et al. Myocardial performance index and biochemical markers

for early detection of doxorubicin-induced cardiotoxicity in children with acute lymphoblastic

leukaemia. Int. J. Clin. Oncol. 2013;18:927–933.

2. Varol E, Akcay S, Ozaydin M, Ozturk O, Cerci SS, Sahin U. Influence of obstructive sleep apnea

on left ventricular mass and global function: sleep apnea and myocardial performance index. Heart

Vessels 2010;25:400–404.

3. Kaya MG, Simsek Z, Sarli B, Buyukoglan H. Myocardial performance index for detection of

subclinical abnormalities in patients with sarcoidosis. J. Thorac. Dis. 2014;6:429–437.

4. Arnlöv J, Ingelsson E, Risérus U, Andrén B, Lind L. Myocardial performance index, a Doppler-

derived index of global left ventricular function, predicts congestive heart failure in elderly men.

Eur. Heart J. 2004;25:2220–2225.

5. Arnlöv J, Lind L, Andrén B, Risérus U, Berglund L, Lithell H. A Doppler-derived index of

combined left ventricular systolic and diastolic function is an independent predictor of

cardiovascular mortality in elderly men. Am. Heart J. 2005;149:902–907.

6. Møller JE, Egstrup K, Køber L, Poulsen SH, Nyvad O, Torp-Pedersen C. Prognostic importance

of systolic and diastolic function after acute myocardial infarction. Am. Heart J. 2003;145:147–153.

7. Biering-Sørensen T, Mogelvang R, Søgaard P, et al. Prognostic Value of Cardiac Time Intervals

by Tissue Doppler Imaging M-Mode in Patients With Acute ST-Segment-Elevation Myocardial

Infarction Treated With Primary Percutaneous Coronary Intervention. Circ. Cardiovasc. Imaging

2013;6:457–465.

17

8. Escande W, Duva Pentiah A, Coisne A, et al. Left ventricular myocardial performance index

predicts poor outcome during COPD exacerbation. Int. J. Cardiol. 2014;173:575–579.

9. Weissler AM, Harris WS, Schoenfeld CD. Systolic time intervals in heart failure in man.

Circulation 1968;37:149–159.

10. Tei C, Ling LH, Hodge DO, et al. New index of combined systolic and diastolic myocardial

performance: a simple and reproducible measure of cardiac function--a study in normals and dilated

cardiomyopathy. J. Cardiol. 1995;26:357–366.

11. Tei C. New non-invasive index for combined systolic and diastolic ventricular function. J.

Cardiol. 1995;26:135–136.

12. Cui W, Roberson DA, Chen Z, Madronero LF, Cuneo BF. Systolic and diastolic time intervals

measured from Doppler tissue imaging: normal values and Z-score tables, and effects of age, heart

rate, and body surface area. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr.

2008;21:361–370.

13. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Novel approach to measure myocardial

performance index: pulsed-wave tissue Doppler echocardiography. Echocardiogr. Mt. Kisco N

2003;20:503–510.

14. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Value of measuring myocardial

performance index by tissue Doppler echocardiography in normal and diseased heart. Jpn. Heart J.

2003;44:403–416.

15. Harada K, Tamura M, Toyono M, Oyama K, Takada G. Assessment of global left ventricular

function by tissue Doppler imaging. Am. J. Cardiol. 2001;88:927–932, A9.

18

16. Voigt J-U, Lindenmeier G, Exner B, et al. Incidence and characteristics of segmental

postsystolic longitudinal shortening in normal, acutely ischemic, and scarred myocardium. J. Am.

Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 2003;16:415–423.

17. Kjaergaard J, Hassager C, Oh JK, Kristensen JH, Berning J, Sogaard P. Measurement of cardiac

time intervals by Doppler tissue M-mode imaging of the anterior mitral leaflet. J. Am. Soc.

Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 2005;18:1058–1065.

18. Biering-Sørensen T, Mogelvang R, Pedersen S, Schnohr P, Sogaard P, Jensen JS. Usefulness of

the myocardial performance index determined by tissue Doppler imaging m-mode for predicting

mortality in the general population. Am. J. Cardiol. 2011;107:478–483.

19. Mogelvang R, Sogaard P, Pedersen SA, et al. Cardiac dysfunction assessed by

echocardiographic tissue Doppler imaging is an independent predictor of mortality in the general

population. Circulation 2009;119:2679–2685.

20. Mogelvang R, Sogaard P, Pedersen SA, Olsen NT, Schnohr P, Jensen JS. Tissue Doppler

echocardiography in persons with hypertension, diabetes, or ischaemic heart disease: the

Copenhagen City Heart Study. Eur. Heart J. 2009;30:731–739.

21. Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification: a report

from the American Society of Echocardiography’s Guidelines and Standards Committee and the

Chamber Quantification Writing Group, developed in conjunction with the European Association of

Echocardiography, a branch of the European Society of Cardiology. J. Am. Soc. Echocardiogr. Off.

Publ. Am. Soc. Echocardiogr. 2005;18:1440–1463.

22. Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and

weight be known. 1916. Nutr. Burbank Los Angel. Cty. Calif 1989;5:303–311; discussion 312–313.

19

23. Dalen H, Thorstensen A, Aase SA, et al. Segmental and global longitudinal strain and strain rate

based on echocardiography of 1266 healthy individuals: the HUNT study in Norway. Eur. J.

Echocardiogr. J. Work. Group Echocardiogr. Eur. Soc. Cardiol. 2010;11:176–183.

24. Chung AK, Das SR, Leonard D, et al. Women have higher left ventricular ejection fractions

than men independent of differences in left ventricular volume: the Dallas Heart Study. Circulation

2006;113:1597–1604.

25. Weidemann F, Jamal F, Sutherland GR, et al. Myocardial function defined by strain rate and

strain during alterations in inotropic states and heart rate. Am. J. Physiol. Heart Circ. Physiol.

2002;283:H792–799.

26. Tei C, Nishimura RA, Seward JB, Tajik AJ. Noninvasive Doppler-derived myocardial

performance index: correlation with simultaneous measurements of cardiac catheterization

measurements. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 1997;10:169–178.

27. Duzenli MA, Ozdemir K, Aygul N, Soylu A, Aygul MU, Gök H. Comparison of myocardial

performance index obtained either by conventional echocardiography or tissue Doppler

echocardiography in healthy subjects and patients with heart failure. Heart Vessels 2009;24:8–15.

28. Meric M, Yesildag O, Yuksel S, et al. Tissue doppler myocardial performance index in patients

with heart failure and its relationship with haemodynamic parameters. Int. J. Cardiovasc. Imaging

2014.

29. Nagueh SF, Appleton CP, Gillebert TC, et al. Recommendations for the evaluation of left

ventricular diastolic function by echocardiography. J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc.

Echocardiogr. 2009;22:107–133.

20

30. Su H-M, Lin T-H, Voon W-C, et al. Correlation of Tei index obtained from tissue Doppler

echocardiography with invasive measurements of left ventricular performance. Echocardiogr. Mt.

Kisco N 2007;24:252–257.

21

Figure legends

Figure 1: Cardiac time intervals assessed by a color tissue Doppler imaging (TDI) M-mode

line through the mitral leaflet.

Left: Four-chamber gray-scale (bottom) and color TDI (top) views in end-systole displaying the

position of the M-mode line used for measuring the cardiac time intervals. Right: Color diagram of

the TDI M-mode line through the mitral leaflet.

MV=Mitral Valve; MVC=MV Closing; AVO=Aortic Valve Opening; AVC=Aortic Valve Closure;

MVO=MV Opening.

Figure 2: The expected cardiac time intervals according to age and gender in participants

without cardiovascular disease or cardiovascular risk factors.

Depicting the expected values of the cardiac time intervals with increasing age for participants

stratified according to gender. Dotted curves indicate 95% confidence intervals. IVCT = Isovolumic

Contraction Time; IVRT = Isovolumic Relaxation Time; ET = Ejection Time; MPI = Myocardial

Performance Index.

Figure 3: The scatter plots for MPITDI and MPIConv according to the values of dP/dt max and

dP/dt min, respectively.

Depicting the scatter plots and the expected values for MPITDI and MPIConv according to the values

of dP/dt max (Figure 3a) and dP/dt min (Figure 3b), respectively. Dotted curves indicate 95%

confidence intervals. dP/dt max= the rate of LV pressure rise in early systole; dP/dt min= the rate of

LV pressure decline in early diastole; MPITDI=Myocardial Performance Index by TDI M-mode;

MPIConv=Myocardial Performance Index by the conventional method.

22

Table 1. Population characteristics in participants without cardiovascular disease or

cardiovascular risk factors (n=)

BMI=Body Mass Index; LVEF=Left Ventricular Ejection Fraction; LVMI=Left Ventricular Mass

Index; LVIDd= left ventricular dimension in end-diastole; LAD=Left atrium diameter in end-

systole; E=peak transmitral early diastolic inflow velocity; A=peak transmitral late diastolic inflow

velocity; DT=deceleration time of early diastolic inflow; e’=average peak early diastolic

longitudinal mitral annular velocity determined by color tissue Doppler imaging.

Women

(n=554)

Men

(n=427)

p-value

Age (years)

50±15 49±14 0.69

Systolic Blood Pressure (mmHg) 118±12 123±10 <0.001

Diastolic Blood Pressure (mmHg) 72±9 75±8 <0.001

Heart Rate (beats per minute) 65±10 64±11 0.038

Cholesterol 5.3±1.1 5.3±1.1 0.93

BMI (kg/m2) 23.7±3.2 25.1±3.0 <0.001

LVEF < 50% 0.2% 0.7% 0.19

LVIDd (cm) 4.6±0.4 5.1±0.5 <0.001

LAD (cm) 3.2±0.3 3.5±0.4 <0.001

E/A ratio < 1 26% 25% 0.74

DT (ms) 158±32 169±40 <0.001

E/e’

8.9 (7.6-10.5) 8.5 (7.4-10.2) 0.06

LVMI 72.4 (64.4-81.7) 88.1 (77.0-99.9) <0.001

23

Table 2. Normal values of the cardiac time intervals in healthy participants stratified according to age category and gender (n=989)

Women

(n=553)

Men

(n=421)

Overall

(n=553)

Age category

20 to 39

(n=150)

Age category

40 to 59

(n=252)

Age category

60 or above

(n=151)

p-value for

age category

difference

Overall

(n=421)

Age category

20 to 39

(n=101)

Age category

40 to 59

(n=210)

Age category

60 or above

(n=110)

p-value for

age category

difference

p-value for

gender

difference

IVRT (ms) 92 (20) 78 (16) 93 (16) 106 (20) <0.001 94 (20) 78 (15) 95 (16) 109 (18) <0.001 0.13

IVCT (ms) 36 (13) 32 (12) 37 (12) 38 (14) <0.001 34 (11) 35 (11) 33 (11) 36 (12) 0.07 0.012*

ET (ms) 293 (21) 291 (19) 296 (20) 289 (23) 0.001 281 (24) 284 (19) 280 (23) 279 (28) 0.34 <0.001*

IVRT/ET 0.32 (0.07) 0.27 (0.05) 0.31 (0.06) 0.37 (0.07) <0.001 0.34 (0.08) 0.28 (0.06) 0.34 (0.06) 0.40 (0.08) <0.001 <0.001*

IVCT/ET 0.12 (0.05) 0.11 (0.04) 0.13 (0.04) 0.13 (0.05) <0.001 0.12 (0.04) 0.13 (0.04) 0.12 (0.04) 0.13 (0.05) 0.10 0.61

MPI 0.44 (0.10) 0.38 (0.08) 0.44 (0.08) 0.51 (0.10) <0.001 0.46 (0.10) 0.40 (0.08) 0.46 (0.09) 0.53 (0.10) <0.001 0.002*

Numbers in parenthesis are standard deviations. BMI=Body Mass Index; LVMI=Left Ventricular Mass Index; LVIDd= left ventricular

dimension in end-diastole; LAD=Left atrium diameter in end-systole; DT=deceleration time of early diastolic inflow; IVCT=Isovolumic

Contraction Time; IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI = Myocardial Performance Index

* remained statistical significantly different between the genders after multivariable adjustment for age, BMI, heart rate, systolic and

diastolic blood pressure, LVMI, LVIDd, LAD, and DT.

24

Table 3. Association between MPITDI and MPIConv, and invasive and echocardiographic measures of systolic and diastolic (n=44)

MPITDI MPIConv

Beta (95% CI) p-value Beta (95% CI) p-value

Invasive measures of systolic and diastolic function:

dP/dt max, per 100 mmHg/sec increase -0.015 (-0.029 to -0.000) 0.047 0.011 (-0.005 to 0.0263) 0.18

dP/dt min, per 100 mmHg/sec decrease 0.012 (-0.003 to 0.027) 0.12 -0.001 (-0.017 to 0.016) 0.94

Conventional and novel echocardiographic measures of systolic and diastolic function:

LVEF, per 1 % increase -0.743 (-1.338 to -0.148) 0.016 -0.613 (-1.220 to -0.007) 0.048

e’, per 1 cm/sec decrease 0.035 (0.018 to 0.053) <0.001 0.021 (0.001 to 0.040) 0.026

E/e’, per 1 increase 0.003 (-0.007 to 0.013) 0.56 0.002 (-0.008 to 0.011) 0.74

GLS (%), per 1 % decrease 0.021 (0.010 to 0.033) 0.001 0.009 (-0.004 to 0.022) 0.16

GL SR s, per 1 sec-1

decrease 0.525 (0.306 to 0.745) <0.001 0.240 (-0.002 to 0.482) 0.051

GL SR e, per 1 sec-1

increase -0.481 (-0.618 to -0.343) <0.001 -0.361 (-0.511 to -0.210) <0.001

dP/dt max= the rate of LV pressure rise in early systole; dP/dt min= the rate of LV pressure decline in early diastole; LVEF=Left

Ventricular Ejection Fraction; E=peak transmitral early diastolic inflow velocity; e’=average peak early diastolic longitudinal mitral

25

annular velocity determined by pulsed-wave tissue Doppler imaging; GLS=global longitudinal strain; GL SR s=global longitudinal systolic

strain rate; GL SR e=global longitudinal early diastolic strain rate; MPITDI=Myocardial Performance Index by TDI M-mode;

MPIConv=Myocardial Performance Index by the conventional method

26

Table 4. Diagnostic accuracy of MPITDI for diagnosing impaired systolic and/or diastolic

function

Cut-offs Sensitivity Specificity PPV NPV

MPITDI

> 0.54 for women

> 0.56 for men

26%

100% 100% 3%

MPITDI = Myocardial Performance Index by TDI M-mode; PPV = Positive Predictive Value; NPV

= Negative Predictive Value.

27

Figure 1

28

Figure 2

60

80

10

01

20

14

0

IVR

T (

ms)

20 40 60 80 100Age (Years)

Women Men

.2.3

.4.5

IVR

T/E

T

20 40 60 80 100Age (Years)

Women Men

.3.4

.5.6

MP

I

20 40 60 80 100Age (Years)

Women Men

30

35

40

45

IVC

T (

ms)

20 40 60 80 100Age (Years)

Women Men

27

02

80

29

03

00

ET

(m

s)

20 40 60 80 100Age (Years)

Women Men

.1.1

2.1

4.1

6

IVC

T/E

T

20 40 60 80 100Age (Years)

Women Men

29

Figure 3a

.2.4

.6.8

1

500 1000 1500 2000 2500dP/dt max (mmHg/sec)

.2.4

.6.8

11.2

500 1000 1500 2000 2500

dP/dt max (mmHg/sec)

MP

I TD

I M

PI C

on

v

R2 = 0.09

R2 = 0.04

30

Figure 3b

.2.4

.6.8

1

-2500 -2000 -1500 -1000dP/dt min (mmHG/sec)

.2.4

.6.8

11.2

-2500 -2000 -1500 -1000

dP/dt min (mmHg/sec)

MP

I TD

I M

PI C

on

v

R2 = 0.00

R2 = 0.06

Peter Godsk Jørgensen and Jan Skov JensenTor Biering-Sørensen, Rasmus Mogelvang, Peter Søgaard, Sune H. Pedersen, Søren Galatius,

Percutaneous Coronary InterventionElevation Myocardial Infarction Treated With Primary−Patients With Acute ST-Segment

Prognostic Value of Cardiac Time Intervals by Tissue Doppler Imaging M-Mode in

Print ISSN: 1941-9651. Online ISSN: 1942-0080 Copyright © 2013 American Heart Association, Inc. All rights reserved.

Dallas, TX 75231is published by the American Heart Association, 7272 Greenville Avenue,Circulation: Cardiovascular Imaging

doi: 10.1161/CIRCIMAGING.112.0002302013;6:457-465; originally published online March 27, 2013;Circ Cardiovasc Imaging. 

http://circimaging.ahajournals.org/content/6/3/457World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circimaging.ahajournals.org//subscriptions/

is online at: Circulation: Cardiovascular Imaging Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer information about this process is available in the

requested is located, click Request Permissions in the middle column of the Web page under Services. FurtherCenter, not the Editorial Office. Once the online version of the published article for which permission is being

can be obtained via RightsLink, a service of the Copyright ClearanceCirculation: Cardiovascular Imagingin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

457

Preservation of normal time intervals is closely associated with normal cardiac biochemistry, mechanics, and

physiology.1 Therefore, cardiac time intervals have been a target of investigation as markers of cardiac dysfunction for decades.2–4 The main concern has been how to obtain these cardiac time intervals in a fast, easy, noninvasive, and reproducible manner. To overcome this problem, Tei et al5 proposed obtaining the intervals from pulsed-wave Doppler echocardiography of the left ventricular (LV) outflow tract and mitral valve (MV) inflow and thereby calculating the index of combined systolic and diastolic performance, the myocardial performance index (MPI).4,5 Tissue Doppler imaging (TDI) is an alternative noninvasive approach to obtain the cardiac time intervals and thereby the MPI.6,7 With this method, the time intervals can be obtained from 1 projection and 1 cardiac cycle. In contrast, with the conventional method described by Tei et al,5 the time intervals are obtained from 2 projections and at least 2 cardiac

cycles4,5 and may therefore be prone to errors caused by heart rate variability, which is not a concern with TDI. However, the time intervals obtained from TDI velocity curves are prone to regional differences in myocardial biochemistry, mechanics, and physiology, which will display regional differences in the cardiac time intervals. This can be overcome by analyzing the global time intervals through evaluating the MV movement by a simple color TDI M-mode analysis8 instead of the regional velocity curves. Thus, using color TDI M-mode through the mitral leaflet to estimate the cardiac time intervals is an improved method reflecting global cardiac time intervals and eliminating beat-to-beat variation. The advantage and validity of this method have previously been demonstrated.9,10

Efforts to improve risk stratification, clarify pathophysi-ological mechanisms, and identify targets for therapeutic

Received October 11, 2012; accepted March 13, 2013.From the Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark (T.B.-S., R.M., P.S., S.H.P., S.G., P.G.J., J.S.J.); and

Clinical Institute of Medicine, Faculty of Health Sciences, University of Copenhagen, Denmark (T.B.-S., P.G.J., J.S.J.).Correspondence to Tor Biering-Sørensen, MD, Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark Niels Andersensvej

65, DK-2900, Post 835, Copenhagen, Denmark. E-mail tor.biering@gmail.com

Background—Color tissue Doppler imaging M-mode through the mitral leaflet is an easy and precise method to estimate all cardiac time intervals from 1 cardiac cycle and thereby obtain the myocardial performance index (MPI). However, the prognostic value of the cardiac time intervals and the MPI assessed by color tissue Doppler imaging M-mode through the mitral leaflet in patients with ST-segment–elevation myocardial infarction (MI) is unknown.

Methods and Results—In total, 391 patients were admitted with an ST-segment–elevation MI, treated with primary percutaneous coronary intervention, and examined by echocardiography a median of 2 days after the ST-segment–elevation MI. Outcome was assessed according to death (n=33), hospitalization with heart failure (n=53), or new MI (n=25). Follow-up time was a median of 25 months. The population was stratified according to tertiles of the MPI. The risk of new MI, being admitted with congestive heart failure or death, increased with increasing tertile of MPI, being ≈3 times as high for the third tertile compared with the first tertile (hazard ratio, 2.8; 95% confidence interval, 1.7–4.7; P<0.001). MPI provided independent prognostic information in a multivariable Cox regression model adjusted for age, sex, previous MI, peak troponin, and systolic and diastolic echocardiographic parameters, with a hazard ratio of 1.24 (P=0.005) for the combined end point per each 0.1 increase in MPI.

Conclusions—MPI assessed by tissue Doppler imaging M-mode is a simple and reproducible measure that provides independent prognostic information, regardless of rhythm, incremental to conventional and novel echocardiographic parameters of systolic and diastolic function in patients with ST-segment–elevation MI treated with primary percutaneous coronary intervention. (Circ Cardiovasc Imaging. 2013;6:457-465.)

Key Words: cardiac time intervals ◼ myocardial performance index ◼ outcome ◼ STEMI ◼ TDI echocardiography

© 2013 American Heart Association, Inc.

Circ Cardiovasc Imaging is available at http://circimaging.ahajournals.org DOI: 10.1161/CIRCIMAGING.112.000230

Original Article

Prognostic Value of Cardiac Time Intervals by Tissue Doppler Imaging M-Mode in Patients With Acute

ST-Segment–Elevation Myocardial Infarction Treated With Primary Percutaneous Coronary InterventionTor Biering-Sørensen, MD; Rasmus Mogelvang, MD, PhD; Peter Søgaard, MD, DMSc; Sune H. Pedersen, MD, PhD; Søren Galatius, MD, DMSc; Peter Godsk Jørgensen, MD;

Jan Skov Jensen, MD, PhD, DMSc

Clinical Perspective on p 465

458 Circ Cardiovasc Imaging May 2013

intervention are of the utmost importance in patients with ST-segment–elevation myocardial infarction (STEMI). Echo-cardiography after an acute MI (AMI) is a routine procedure for risk stratification. The systole and the diastole are interde-pendent and coherent11,12; therefore, improved risk stratifica-tion by combining the systolic and diastolic performance in 1 index may be possible. The aim of this study was to evaluate the prognostic value of the cardiac time intervals and the combined index of systolic and diastolic performance, the MPI, obtained by TDI M-mode method in patients with STEMI treated with primary percutaneous coronary intervention (pPCI).

MethodsStudy PopulationFrom September 2006 to December 2008, a total of 391 patients were admitted with STEMI, treated with pPCI, and underwent a detailed echocardiographic examination at Gentofte Hospital, University of Copenhagen, Copenhagen, Denmark. Five patients were excluded because of inadequate quality of the echocardiographic examination with regard to obtaining the cardiac time intervals.

The definition of an STEMI was presence of chest pain for >30 minutes and <12 hours, persistent ST-segment elevation ≥2 mm in at least 2 contiguous precordial ECG leads or ≥1 mm in at least 2 contiguous limb ECG leads (or a newly developed left bundle-branch block), combined with a troponin I increase >0.5 μg/L.

Baseline data were collected prospectively. Hypertension was de-fined as use of blood pressure–lowering drugs at admission. Diabetes mellitus was defined as fasting plasma glucose concentration ≥7 mmol/L, nonfasting plasma glucose concentration ≥11.1 mmol/L, or the use of glucose-lowering drugs at admission.

Troponin I was measured immediately on admission and after 6 and 12 hours.

EchocardiographyEchocardiography was performed using Vivid 7 ultrasound systems (GE Healthcare, Horten, Norway) with a 3.5-MHz transducer by experienced sonographers. At our institution, echocardiography is performed during the initial 5 days of admission (median, 2 days; interquartile range, 1–3 days). All participants were examined with conventional 2-dimensional echocardiography and pulsed-wave and color TDI according to standardized protocols. All echocardiograms were stored digitally and analyzed offline (EchoPac, GE Healthcare,

Horten, Norway) by a single investigator who was blinded to all other patient data.

Conventional EchocardiographyFrom the parasternal long-axis view, the LV diameters and wall thickness were measured. From the apical position, peak velocity of early and atrial diastolic filling and deceleration time of the E wave were measured using pulsed-wave Doppler to record the mitral inflow. LV ejection fraction (LVEF) was obtained using the modified biplane Simpson method. Left atrial volume was estimated by the area-length method, and the left atrial volume index was calculated. The LV mass index was calculated by the formula using LV linear dimensions.13

Two-Dimensional Strain EchocardiographyTwo-dimensional strain analysis was performed from the apical 4-chamber, 2-chamber, and apical long-axis views (mean, 86 frames per second; SD, 23 frames per second). Peak longitudinal systolic strain was measured in all 3 apical projections and was averaged to provide global estimates.

Tissue Doppler ImagingPulsed-wave TDI tracings were obtained with the range gate placed at the septal and lateral mitral annular segments in the 4-chamber view. The peak longitudinal early diastolic (e') velocity was measured, and the average was calculated from the lateral and septal velocities and used to obtain the E/e'.

Color TDI loops were obtained in the apical 4-chamber view at the highest possible frame rate (mean, 169 frames per second; SD, 33 frames per second). The cardiac time intervals were obtained by placing a 2- to 4-cm straight M-mode line through the septal half of the mitral leaflet in the color TDI 4-chamber view, and the time inter-vals were measured directly from the color diagram (Figure 1). The isovolumic contraction time (IVCT) was defined as the time interval from the MV closure, determined by the color shift from blue/tur-quoise to red at end diastole, to the aortic valve opening, determined by the color shift from blue to red (Figure 1). The ejection time (ET) was defined as the time interval from the aortic valve opening to the aortic valve closing, determined by the color shift from red to blue at end systole (Figure 1). The isovolumic relaxation time (IVRT) was defined as the time interval from the aortic valve closing to the MV opening, determined by the color shift from red-orange to yel-low (Figure 1). The method has previously been validated.9,10 Both isovolumic time intervals were divided by ET creating IVRT/ET and IVCT/ET, respectively, and MPI was calculated as the sum of the 2 ([IVRT+IVCT]/ET).

Figure 1. The cardiac time intervals assessed by a color tissue Doppler imaging (TDI) M-mode line through the mitral leaflet. Left, Four-chamber gray-scale (bottom) and color TDI (top) views in end systole displaying the position of the M-mode line used for measuring the cardiac time intervals. Right, Color diagram of the TDI M-mode line through the mitral leaflet. AVC indicates aortic valve closure; AVO, aortic valve opening; MVC, mitral valve closing; and MVO, mitral valve opening.

Biering-Sørensen et al Cardiac Time Intervals by TDI Predict Outcome After Primary PCI 459

Diastolic function was assessed using mitral inflow velocity pro-files and pulsed-wave TDI tracings from the septal and lateral mitral annulus, according to the present guidelines.14

Patients with atrial fibrillation were defined as having diastolic dysfunction and were graded as 1, 2, or 3 according to the value of E/e'.15

pPCI ProcedurepPCI was performed according to contemporary interventional guide-lines using pretreatment with 300 mg acetylsalicylic acid, 600 mg clopidogrel, and 10 000 IU unfractionated heparin. Glycoprotein IIb/IIIa inhibitors were used at the discretion of the operator. Multivessel disease was defined as 2- or 3-vessel disease and complex lesions as type C lesions. Subsequent medical treatment included anti-ischemic, lipid-lowering, and antithrombotic drugs.

Follow-up and End PointsThe primary end point was the combined end point of all-cause mor-tality, a new MI (re-MI), or admission with congestive heart failure (CHF). Secondary end points were all of the above analyzed sepa-rately. Follow-up was 100%. Follow-up data on re-MI and admis-sion with CHF were obtained from the Danish National Board of Health’s National Patient Registry using International Classification of Diseases, 10th Revision (ICD-10) codes and thoroughly validated using hospital source data. Follow-up data on mortality were collect-ed from the National Person Identification Registry.

StatisticsIn Tables 1 and 2, proportions were compared using the χ2 test, con-tinuous gaussian-distributed variables with the Student t test, and non–gaussian-distributed variables with the Mann–Whitney test. Association with MPI was tested for LVEF and e' by multivariable regression analyses, including the prespecified baseline variables age and sex.

Cumulative survival curves were established by the Kaplan-Meier method and compared by the log-rank test. Hazard ratios (HRs) were calculated by Cox proportional hazards regression analyses. The assumptions of linearity and proportional hazards in the models were tested. First-order interactions between all cardiac time intervals, including the combined indexes, and atrial fibrillation for the combined end point were tested.

Receiver-operating characteristic curves were constructed for IVCT/ET, IVRT/ET, and MPI in effort to find the optimal cutoff value with the highest sensitivity and specificity for predicting the risk of the combined end point.

Intraobserver and interobserver variabilities were determined by Bland-Altman plots and coefficients of variation (CVs) of repeated analyses of stored echocardiographic loops (n=25 for patients with sinus rhythm and n=14 for patients with atrial fibrillation). A value of P≤0.05 in 2-sided tests was considered statistically significant. SPSS for Windows version 20.0 (SPSS Inc, Chicago, IL) was used.

The study was approved by the local scientific ethics committee and the Danish Data Protection Agency. Furthermore, it complied with the second Declaration of Helsinki, and all participants signed informed consent.

ResultsDuring follow-up (median, 25 months; interquartile range, 19–32 months), 25 patients (6.5%) were admitted with a re-MI, 53 (13.7%) were admitted to hospital for CHF, and 33 patients (8.5%) died. The combined end point was reached by 96 patients (24.9%).

Baseline characteristics are displayed in Tables 1 and 2.

MPI and Conventional EchocardiographyThe association between MPI and the established echocardio-graphic parameters of diastolic and systolic function, e' and

LVEF, was tested (Figure 2). MPI increased independently with decreasing values of e' (P<0.001) and LVEF (P<0.001), after adjustment for age and sex.

Cardiac Time Intervals and PrognosisThe risk of a subsequent re-MI, being admitted with CHF or death, increased with increasing tertile of the IVCT (Figure 3A), being 2 times as high in the third tertile com-pared with the first tertile (HR, 2.0; 95% confidence interval [CI], 1.2–3.3; P=0.007). When interpreting the Kaplan-Meier curves, neither the IVRT nor the ET seems to provide eas-ily comprehensible prognostic information because the risk of

Table 1. Baseline Clinical Characteristics in Patients With ST-Segment–Elevation Myocardial Infarction Treated by pPCI Stratified According to Major Adverse Outcome

No Major Adverse Outcome (n=290)

Major Adverse Outcome (n=96) P Value

Age, y* 61±11 67±12 <0.001

Male sex 76% 72% 0.40

MAP, mm Hg* 100±18 102±22 0.32

Heart rate, beats per minute*

76±39 81±25 0.26

Hypertension 32% 33% 0.77

Diabetes mellitus 8% 9% 0.66

Current smoker 52% 47% 0.35

Hypercholesterolemia 17% 19% 0.68

Previous MI 3% 8% 0.030

BMI, kg/m2* 26.9±4.2 25.7±4.9 0.017

Peak troponin I, µg/L † 89 (26–213) 163 (48–310) 0.002

eGFR, mL/min* 74±21 72±24 0.44

Symptom-to-balloon time, min†

180 (120–305) 201 (145–340) 0.10

Complex lesion 44% 53% 0.12

Multivessel disease 27% 31% 0.41

LAD-lesion 45% 48% 0.60

Glycoprotein IIb/IIIa inhibitor

22% 29% 0.16

TIMI grade flow before pPCI

0.61

TIMI 0 flow 62% 66%

TIMI 1 flow 14% 10%

TIMI 2 flow 10% 13%

TIMI 3 flow 15% 12%

TIMI grade flow after pPCI 0.13

TIMI 0 flow 4% 3%

TIMI 1 flow 4% 10%

TIMI 2 flow 8% 10%

TIMI 3 flow 84% 78%

Atrial fibrillation/flutter 2% 8% 0.004

BMI indicates body mass index; eGFR, estimated glomerular filtration rate; LAD, left anterior descending coronary artery; MAP, mean arterial blood pressure; MI, myocardial infarction; pPCI, primary percutaneous coronary intervention; and TIMI, thrombolysis in myocardial infarction classification.

*Mean ±SD.†Median (interquartile range).

460 Circ Cardiovasc Imaging May 2013

the combined end point did not increase incrementally with increasing or decreasing tertile, respectively (Figure 3B and 3C). We found no interactions between cardiac time intervals (including the combined indexes) and atrial fibrillation for the combined end point.

MPI and PrognosisAll the indexes obtained by combining the cardiac time inter-vals IVRT/ET, IVCT/ET, and MPI were associated with an

increased risk of reaching the combined end point with increasing tertile (Figure 4). The risk of a re-MI, being admit-ted with CHF or death, increased with increasing tertile of the combined indexes (Figure 4), being ≈2 times as high for the IVRT/ET and ≈3 times as high for the IVCT/ET and MPI in the third tertile compared with the first tertile (IVRT/ET: HR, 1.9; 95% CI, 1.2–3.2; P=0.010; IVCT/ET: HR, 2.8; 95% CI, 1.6–4.7; P<0.001; MPI: HR, 2.8; 95% CI, 1.7–4.7; P<0.001). Consequently, all the combined indexes were tested as pre-dictors of an adverse outcome using Cox regression models (Table 3). Only the IVRT/ET and the MPI remained indepen-dent predictors of the combined outcome after adjustment for age, sex, peak troponin I, previous AMI, LVEF, peak global longitudinal systolic strain, diastolic function grade, LV internal diameter in diastole/body surface area, and LV mass index (Table 3). The same result was achieved when adjust-ing for the presence of atrial fibrillation, heart rate, mean arterial blood pressure, body mass index, and e', but to main-tain a robust model, these variables were not included in the final multivariable Cox model (Table 3). The optimal cutoff for predicting the combined end point was determined from the receiver-operating characteristic curves and was 0.14 for IVCT/ET, 0.42 for IVRT/ET, and 0.52 for MPI.

MPI in Atrial Fibrillation and OutcomeEven in patients with atrial fibrillation, the MPI was signifi-cantly higher in patients with an adverse outcome (0.45 [95% CI, 0.34–0.55] versus 0.67 [95% CI, 0.58–0.76]; P=0.005; Figure 5). Even after adjustment for age, sex, and LVEF, the MPI remained significantly higher in the group with an adverse outcome (0.44 [95% CI, 0.29–0.58] versus 0.67 [95% CI, 0.56–0.79]; P=0.026).

Intraobserver and Interobserver Variability AnalysisBland-Altman plots of intraobserver and interobserver dif-ferences of MPI are shown in Figure 6. The intraobserver

Table 2. Baseline Echocardiographic Characteristics in Patients With ST-Segment–Elevation Myocardial Infarction Treated by Primary Percutaneous Coronary Intervention Stratified According to Major Adverse Outcome

No Major Adverse Outcome (n=290)

Major Adverse Outcome (n=96) P Value

LVEF, %* 47±8 41±9 <0.001

GLS, %* −12.8±3.7 −10.4±3.5 <0.001

LVIDd/BSA, cm/m2* 2.4 (2.3–2.7) 2.6 (2.4–2.9) <0.001

LVEDV/BSA, mL/m2† 48 (40–58) 50 (42–57) 0.23

LVESV/BSA, mL/m2† 25 (21–31) 29 (22–37) 0.001

LVMI, g/m2† 88 (73–106) 102 (84–118) <0.001

LAVI, mL/m2* 25±7 25±7 0.81

E, m/s* 0.77±0.18 0.77±0.24 0.83

A, m/s* 0.74±0.19 0.75±0.23 0.68

E/A ratio* 1.09±0.37 1.09±0.37 0.95

DT, ms* 199±59 201±75 0.80

e’* 7.6±2.2 6.7±1.9 <0.001

E/e’† 10.3 (8.1–12.4) 11.4 (9.1–14.4) 0.002

Grade of diastolic function

<0.001

Normal 32% 15%

Grade I dysfunction 23% 24%

Grade II dysfunction 40% 47%

Grade III dysfunction 4% 15%

Cardiac time intervals

IVRT, ms* 101±21 105±28 0.31

IVCT, ms* 30±14 35±14 0.004

ET, ms* 258±30 241±42 <0.001

IVRT/ET* 0.40±0.10 0.44±13 <0.001

IVCT/ET* 0.12±0.06 0.15±0.07 <0.001

MPI* 0.52±0.13 0.59±0.16 <0.001

A indicates peak transmitral late diastolic inflow velocity; BSA, body surface area; DT, deceleration time of early diastolic inflow; E, peak transmitral early diastolic inflow velocity; e’, average peak early diastolic longitudinal mitral annular velocity determined by pulsed-wave TDI; ET, ejection time; GLS, peak global longitudinal systolic strain; IVCT, isovolumic contraction time; IVRT, isovolumic relaxation time; LAVI, left atrial volume index; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVIDd, left ventricular internal diameter in diastole; LVMI, left ventricular mass index; and MPI, myocardial performance index.

*Mean ±SD.†Median (interquartile range).

Figure 2. Association between the myocardial performance index (MPI) and established echocardiographic parameters of diastolic and systolic function. Depicted is the association between the MPI and the ejection fraction (EF) and average peak early diastolic longitudinal mitral annular velocity determined by pulsed-wave tissue Doppler imaging (e'). Error bars represents SEs.

Biering-Sørensen et al Cardiac Time Intervals by TDI Predict Outcome After Primary PCI 461

variability analysis showed a mean±SD difference of −0.004±0.029 (CV, 6%) for patients with sinus rhythm and 0.031±0.055 (CV, 9%) for patients with atrial fibrillation (Figure 6A). The interobserver variability analysis showed a mean±SD difference of −0.022±0.084 (CV, 16%) for patients with sinus rhythm and 0.017±0.096 (CV, 16%) for patients with atrial fibrillation (Figure 6B).

DiscussionThis prospective study of STEMI patients illustrates the prog-nostic information that can be obtained from cardiac time intervals and emphasizes the importance of combining the evaluation of the systolic and diastolic performance. Our study is the first to demonstrate that the cardiac time intervals can be valuable in risk stratification of patients with atrial fibrillation.

Figure 3. Cardiac time intervals and outcome. Kaplan-Meier curves depicting cumulative probability of staying event free for patients stratified into tertiles of the isovolumic contraction time (IVCT; first tertile <24 milliseconds; second tertile ≥24–<34 mil-liseconds; third tertile ≥34 milliseconds; A), of the isovolumic relaxation time (IVRT; first tertile <93 milliseconds; second tertile ≥93–<110 milliseconds; third tertile ≥110 milliseconds; B), and of the ejection time (ET; first tertile <244 milliseconds; second tertile ≥244–<271 milliseconds; third tertile ≥271 milliseconds; C). pPCI indicates primary percutaneous coronary intervention.

Figure 4. The combined indexes of the cardiac time intervals and outcome. Kaplan-Meier curves depicting cumulative prob-ability of staying event-free for patients stratified into tertiles of the combined index isovolumic contraction time (IVCT)/ejection time (ET) (first tertile <0.09; second tertile ≥0.09–<0.14; third ter-tile ≥0.14; A), of the combined index isovolumic relaxation time (IVRT)/ET (first tertile <0.36; second tertile ≥0.36–<0.43; third ter-tile ≥0.43; B), and of the myocardial performance index (MPI; first tertile <0.46; second tertile ≥0.46–<0.57; third tertile ≥0.57; C). pPCI indicates primary percutaneous coronary intervention.

462 Circ Cardiovasc Imaging May 2013

We found that MPI was associated with established echo-cardiographic parameters of both diastolic and systolic function. MPI increased with decreasing systolic function, determined by LVEF, and increased with decreasing diastolic function, determined by e' (Figure 2). Furthermore, prolon-gation of the systolic time interval, IVCT, and higher values of all the combined indexes, IVCT/ET, IVRT/ET, and MPI, were associated with an adverse outcome (Table 3 and Figures 3 and 4). In addition, the combined indexes of systolic and diastolic performance, the IVRT/ET, and the MPI provided independent prognostic information incremental to conven-tional and novel echocardiographic parameters of systolic and

diastolic function and all other strong predictors in our popu-lation (Table 3).

MPI and Conventional EchocardiographyInvasive measures of systolic and diastolic performance, the LVEF and LV relaxation constant (τ), have previously been illustrated to be independent predictors of MPI obtained from TDI velocity curves.16 We found LVEF and e' to be indepen-dent predictors of MPI determined by color TDI M-mode through the mitral leaflet. This illustrates that an ailing sys-tolic or diastolic performance can be detected by an increas-ing value of the MPI when assessed by TDI. These results are interesting when we consider the potential advantages of obtaining the cardiac time intervals by color TDI M-mode through the mitral leaflet compared with the conventional method.4,5 Besides obtaining all the cardiac time intervals from 1 cardiac cycle, the MPI obtained by TDI M-mode is less influenced by physical parameters compared with the conventional method.10 Hence, MPI based on simple cardiac mechanisms, that is, the passively opening and closing of the MV, is less confounded than MPI based on calculations from versatile blood flow patterns through the mitral and aorta valves. Furthermore, the precision and reproducibility of the cardiac time intervals and MPI are improved when they are obtained by the TDI M-mode method compared with the conventional method.9,10 In addition, the time intervals achieved by clear color shifts marking minimal changes in direction in the MV by the color TDI M-mode method are very easy to identify (Figure 1). In contrast, the time inter-vals obtained by the conventional method4,5 are assessed from velocity curves in which the signals may be scattered, making

Figure 5. Myocardial performance index (MPI) in patients with atrial fibrillation stratified according to outcome. Circles indicate absolute values; dotted lines indicate means.

Table 3. Unadjusted and Adjusted Cox Proportional Hazards Regression Models Depicting the Combined Indexes of the Cardiac Time Intervals as Predictors of Outcome

Combined End Point (96 Events) CHF (53 Events) re-MI (25 Events) Mortality (33 Events)

Hazard Ratio (95% CI) P Value

Hazard Ratio (95% CI) P Value

Hazard Ratio (95% CI) P Value

Hazard Ratio (95% CI) P Value

Unadjusted model

IVRT/ET per 0.1 increase 1.34 (1.14–1.57) <0.001 1.28 (1.03–1.59) 0.028 1.40 (1.04–1.89) 0.025 1.12 (0.84–1.50) 0.44

IVCT/ET per 0.1 increase 1.53 (1.26–1.86) <0.001 1.68 (1.34–2.11) <0.001 1.39 (0.94–2.08) 0.10 1.41 (0.99–2.00) 0.06

MPI per 0.1 increase 1.30 (1.16–1.46) <0.001 1.32 (1.14–1.54) <0.001 1.31 (1.05–1.63) 0.015 1.17 (0.95–1.43) 0.15

Multivariable model adjusted for age, sex, peak TnI, pre-MI, LVEF, diastolic function grade, LVMI, and LVIDd/BSA

IVRT/ET per 0.1 increase 1.29 (1.08–1.54) 0.005 1.21 (0.94–1.56) 0.15 1.18 (0.81–1.72) 0.40 1.20 (0.89–1.60) 0.23

IVCT/ET per 0.1 increase 1.42 (1.00–2.02) 0.053 1.43 (0.90–2.26) 0.13 1.34 (0.63–2.89) 0.45 1.42 (0.77–2.60) 0.26

MPI per 0.1 increase 1.26 (1.09–1.46) 0.002 1.22 (0.99–1.50) 0.06 1.17 (0.86–1.60) 0.32 1.20 (0.94–1.52) 0.15

Multivariable model adjusted for age, sex, peak TnI, pre-MI, GLS, diastolic function grade, LVMI, and LVIDd/BSA

IVRT/ET per 0.1 increase 1.30 (1.08–1.56) 0.005 1.22 (0.94–1.58) 0.14 1.26 (0.86–1.84) 0.23 1.20 (0.88–1.64) 0.26

IVCT/ET per 0.1 increase 1.43 (0.98–2.08) 0.06 1.53 (0.93–2.49) 0.09 1.54 (0.70–3.38) 0.28 1.16 (0.59–2.28) 0.66

MPI per 0.1 increase 1.27 (1.09–1.48) 0.002 1.25 (1.00–1.55) 0.046 1.25 (0.92–1.71) 0.16 1.16 (0.89–1.52) 0.26

Multivariable model adjusted for age, sex, peak TnI, pre-MI, LVEF, GLS, diastolic function grade, LVMI, and LVIDd/BSA

IVRT/ET per 0.1 increase 1.28 (1.07–1.53) 0.008 1.22 (0.94–1.57) 0.14 1.14 (0.77–1.68) 0.51 1.18 (0.87–1.61) 0.29

IVCT/ET per 0.1 increase 1.30 (0.89–1.90) 0.18 1.36 (0.83–2.24) 0.23 1.25 (0.56–2.75) 0.59 1.13 (0.58–2.20) 0.72

MPI per 0.1 increase 1.24 (1.07–1.45) 0.005 1.22 (0.98–1.51) 0.08 1.14 (0.82–1.57) 0.44 1.15 (0.88–1.49) 0.30

BSA indicates body surface area; CHF, congestive heart failure; CI, confidence interval; ET, ejection time; GLS, peak global longitudinal systolic strain; IVCT, isovolumic contraction time; IVRT, isovolumic relaxation time; LVEF, left ventricular ejection fraction; LVIDd, left ventricular internal diameter in diastole; LVMI, left ventricular mass index; MI, myocardial infarction; MPI, myocardial performance index; peak TnI, peak troponin I; and pre-MI, previous acute myocardial infarction.

Biering-Sørensen et al Cardiac Time Intervals by TDI Predict Outcome After Primary PCI 463

it hard to accurately define the cardiac time intervals. Further-more, when good imaging quality is difficult to obtain (eg, in patients shortly after a pPCI who are not mobilizable), it is often possible to visualize the MV in the apical view and to asses its longitudinal movement by color TDI M-mode.

Cardiac Time Interval, MPI, and PrognosisThe MPI obtained by the conventional method can provide prognostic information in patients with various cardiac condi-tions,17–20 including AMI.21–25 Our study is the first to evalu-ate the prognostic value of the MPI assessed by TDI M-mode through the mitral leaflet in an STEMI population treated by pPCI. Previous studies on the predictive power of the MPI in patients with AMI have included patients with both non-STEMI and STEMI treated with either thrombolytic therapy or PCI. Our population is more homogeneous and less con-founded by difference in therapy.

We found that the cardiac time intervals when evaluated separately provided ambiguous and not easily comprehensible prognostic information (Figure 3). Only the IVCT seemed to provide prognostic information when interpreting the Kaplan-Meier curves, depicting the proportion of patients staying event free for patients stratified into tertiles of the cardiac

time intervals (Figure 3A). For the IVRT and ET, the risk of the combined end point did not increase incrementally with increasing or decreasing tertile, respectively (Figure 3B and 3C). However, when we corrected the IVCT and the IVRT for heart rate by dividing it by ET,16 both combined indexes provided incremental prognostic information with increasing tertile (Figure 4A and 4B). This result is interesting because neither the IVRT nor the ET provided incremental prognostic information when evaluating them separately (Figure 3B and 3C), but when the information about the systolic and the dia-stolic performance is combined in 1 index (and thereby also indirectly adjusted for heart rate), the prognostic information becomes evident (Figure 4A and 4B). Only the combined indexes, including information on both the systolic and dia-stolic performance (the MPI and the IVRT/ET), were strong prognosticators after multivariable adjustment for the conven-tional and novel echocardiographic parameters of systolic and diastolic function (Table 3). The index including information only on the systolic performance, the IVCT/ET, did not pro-vide prognostic information independent of other echocardio-graphic parameters (Table 3). Furthermore, we propose using a cutoff of 0.42 for IVRT/ET and 0.52 for MPI when these indexes are used for risk stratification in STEMI patients.

Figure 6. Intraobserver and interobserver variability analysis. Bland-Altman plots of intraobserver (A) and interobserver (B) differences of myocardial performance index assessed by color tissue Doppler imaging M-mode through the mitral leaflet for patients with sinus rhythm (n=25) and for patients with atrial fibrillation (n=14) showing the mean difference (solid lines) and 95% limits of agreement (dotted lines).

464 Circ Cardiovasc Imaging May 2013

The MPI has previously been demonstrated to be a supe-rior predictor of cardiovascular mortality and CHF compared with conventional echocardiographic measures of systolic and diastolic performance in a population of elderly men.17,18 In addition, this superiority of the predictive capacity of the MPI compared with conventional echocardiographic measures has been demonstrated in an AMI population.22 These and our results emphasize the prognostic information gained by combining the interdependent and coherent relations of sys-tolic and diastolic function in a simple and feasible index. In addition, in accordance with our results, we have previously demonstrated that MPI obtained by the M-mode method was an independent predictor of all-cause mortality in the general population, providing prognostic information incremental to known confounders such as age, sex, body mass index, heart rate, blood pressure, and ischemic heart disease. In contrast, the MPI obtained by the conventional method was not an inde-pendent predictor when adjusted for the same confounders.10 The discrepancy in the predictive capacity of the MPI when obtained by the 2 different methods is probably attributable to the fact that the method by Tei et al4,5 seems to be influenced by more physical parameters than the M-mode method.10 This makes the MPI obtained by TDI M-mode less complex to interpret in a clinical setting.

MPI in Atrial Fibrillation and OutcomeIn addition, it is not possible to obtain cardiac time inter-vals from velocity curves in patients with atrial fibrillation, regardless of whether they are obtained by the conventional method4,5 or by the newer TDI method.6,7 This is attributable to the absence of the A and a' velocity curves in patients with atrial fibrillation, which are needed to obtain the time intervals by both methods. In contrast, all cardiac time intervals can be obtained by the color TDI M-mode method, regardless of the presence of atrial fibrillation. Thus, in patients with atrial fibrillation, color M-mode MPI can differentiate between high- and low-risk patients (Figure 5), even after adjustment for age, sex, and LVEF. However, the reproducibility seems lower in patients with atrial fibrillation than in patients with sinus rhythm (Figure 6). Nevertheless, the intraobserver and interobserver variability for TDI M-mode MPI in patients with atrial fibrillation seems superior to the variability found for the conventional method described by Tei et al5 in patients with sinus rhythm,10 which provides more optimism for this new method.

LimitationsThe risk of residual confounders always exists in a nonran-domized study. Furthermore, we did not investigate the cause of death. However, we assume that individuals with cardiac dysfunction determined by echocardiography after STEMI are more likely to die of cardiovascular causes and that, if we were able to limit our analysis to cardiovascular deaths, the prognostic impact of MPI would be even greater.

The investigator was not blinded for other echocardiographic parameters, which theoretically could have a potential for bias. Nevertheless, the combined indexes provide independent prog-nostic information after adjustment for all other systolic and diastolic parameters in our population, which would not be the

case if the combined indexes contained only prognostic infor-mation gained from other echocardiographic parameters.

Echocardiography was performed a median of 2 days after pPCI, which is the typical time span for echocardiographic risk assessment in east Denmark, where various degrees of myocardial stunning might be present and may impact the cardiac time intervals. However, myocardial function can be affected for as long as 2 weeks after the intervention,26 at which point the myocardial function could be influenced by long-term compensatory mechanisms and the effects of addi-tional therapeutic interventions, so the optimal time for risk assessment after pPCI is still to be investigated.

ConclusionsMPI assessed by TDI M-mode is a simple and reproducible measure that provides independent prognostic information, regardless of rhythm, incremental to conventional and novel echocardiographic parameters of systolic and diastolic func-tion in patients with STEMI treated with pPCI.

Sources of FundingThis study was financially supported by the Faculty of Health Sciences, University of Copenhagen, Denmark. The sponsor had no role in the study design, data collection, data analysis, data interpreta-tion, or writing of the article.

DisclosuresNone.

References 1. Oh JK, Tajik J. The return of cardiac time intervals: the phoenix is rising.

J Am Coll Cardiol. 2003;42:1471–1474. 2. Weissler AM, Harris WS, Schoenfeld CD. Systolic time intervals in heart

failure in man. Circulation. 1968;37:149–159. 3. Mancini GB, Costello D, Bhargava V, Lew W, LeWinter M, Karliner JS.

The isovolumic index: a new noninvasive approach to the assessment of left ventricular function in man. Am J Cardiol. 1982;50:1401–1408.

4. Tei C. New non-invasive index for combined systolic and diastolic ven-tricular function. J Cardiol. 1995;26:135–136.

5. Tei C, Ling LH, Hodge DO, Bailey KR, Oh JK, Rodeheffer RJ, Tajik AJ, Seward JB. New index of combined systolic and diastolic myocardial per-formance: a simple and reproducible measure of cardiac function–a study in normals and dilated cardiomyopathy. J Cardiol. 1995;26:357–366.

6. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Novel approach to measure myocardial performance index: pulsed-wave tissue Doppler echocardiography. Echocardiography. 2003;20:503–510.

7. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Value of measur-ing myocardial performance index by tissue Doppler echocardiography in normal and diseased heart. Jpn Heart J. 2003;44:403–416.

8. Voigt JU, Lindenmeier G, Exner B, Regenfus M, Werner D, Reulbach U, Nixdorff U, Flachskampf FA, Daniel WG. Incidence and characteris-tics of segmental postsystolic longitudinal shortening in normal, acutely ischemic, and scarred myocardium. J Am Soc Echocardiogr. 2003;16: 415–423.

9. Kjaergaard J, Hassager C, Oh JK, Kristensen JH, Berning J, Sogaard P. Measurement of cardiac time intervals by Doppler tissue M-mode imaging of the anterior mitral leaflet. J Am Soc Echocardiogr. 2005;18:1058–1065.

10. Biering-Sørensen T, Mogelvang R, Pedersen S, Schnohr P, Sogaard P, Jensen JS. Usefulness of the myocardial performance index determined by tissue Doppler imaging m-mode for predicting mortality in the general population. Am J Cardiol. 2011;107:478–483.

11. Claessens TE, De Sutter J, Vanhercke D, Segers P, Verdonck PR. New echocardiographic applications for assessing global left ventricular dia-stolic function. Ultrasound Med Biol. 2007;33:823–841.

12. Yip GW, Zhang Y, Tan PY, Wang M, Ho PY, Brodin LA, Sanderson JE. Left ventricular long-axis changes in early diastole and systole: impact of systolic function on diastole. Clin Sci (Lond). 2002;102:515–522.

Biering-Sørensen et al Cardiac Time Intervals by TDI Predict Outcome After Primary PCI 465

13. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MSJ, Stewart WJ. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr. 2005;18:1440–1463.

14. Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth OA, Waggoner AD, Flachskampf FA, Pellikka PA, Evangelista A. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. J Am Soc Echocardiogr. 2009;22:107–133.

15. Kosiuk J, Van Belle Y, Bode K, Kornej J, Arya A, Rolf S, Husser D, Hindricks G, Bollmann A. Left ventricular diastolic dysfunction in atrial fibrillation: predictors and relation with symptom severity. J Cardiovasc Electrophysiol. 2012;23:1073–1077.

16. Su HM, Lin TH, Voon WC, Lee KT, Chu CS, Yen HW, Lai WT, Sheu SH. Correlation of Tei index obtained from tissue Doppler echocar-diography with invasive measurements of left ventricular performance. Echocardiography. 2007;24:252–257.

17. Arnlöv J, Ingelsson E, Risérus U, Andrén B, Lind L. Myocardial per-formance index, a Doppler-derived index of global left ventricular function, predicts congestive heart failure in elderly men. Eur Heart J. 2004;25:2220–2225.

18. Arnlöv J, Lind L, Andrén B, Risérus U, Berglund L, Lithell H. A Doppler-derived index of combined left ventricular systolic and diastolic function is an independent predictor of cardiovascular mortality in elderly men. Am Heart J. 2005;149:902–907.

19. Dujardin KS, Tei C, Yeo TC, Hodge DO, Rossi A, Seward JB. Prognostic value of a Doppler index combining systolic and diastolic performance in idiopathic-dilated cardiomyopathy. Am J Cardiol. 1998;82:1071–1076.

20. Kim H, Yoon HJ, Park HS, Cho YK, Nam CW, Hur SH, Kim YN, Kim KB. Usefulness of tissue Doppler imaging-myocardial performance index in the evaluation of diastolic dysfunction and heart failure with preserved ejection fraction. Clin Cardiol. 2011;34:494–499.

21. Møller JE, Søndergaard E, Poulsen SH, Egstrup K. The Doppler echocardio-graphic myocardial performance index predicts left-ventricular dilation and cardiac death after myocardial infarction. Cardiology. 2001;95:105–111.

22. Møller JE, Egstrup K, Køber L, Poulsen SH, Nyvad O, Torp-Pedersen C. Prognostic importance of systolic and diastolic function after acute myo-cardial infarction. Am Heart J. 2003;145:147–153.

23. Poulsen SH, Jensen SE, Nielsen JC, Møller JE, Egstrup K. Serial changes and prognostic implications of a Doppler-derived index of combined left ventricular systolic and diastolic myocardial performance in acute myo-cardial infarction. Am J Cardiol. 2000;85:19–25.

24. Ascione L, De Michele M, Accadia M, Rumolo S, Damiano L, D’Andrea A, Guarini P, Tuccillo B. Myocardial global performance index as a pre-dictor of in-hospital cardiac events in patients with first myocardial infarc-tion. J Am Soc Echocardiogr. 2003;16:1019–1023.

25. Souza LP, Campos O, Peres CA, Machado CV, Carvalho AC. Echocardiographic predictors of early in-hospital heart failure during first ST-elevation acute myocardial infarction: does myocardial performance index and left atrial volume improve diagnosis over conventional param-eters of left ventricular function? Cardiovasc Ultrasound. 2011;9:17.

26. Braunwald E, Kloner RA. The stunned myocardium: prolonged, postisch-emic ventricular dysfunction. Circulation. 1982;66:1146–1149.

CLINICAL PERSPECTIVEChanges in cardiac time intervals are sensitive markers of an ailing myocardium. As left ventricular (LV) systolic function deteriorates, the time it takes for myocardial myocytes to achieve an LV pressure equal to that of aorta increases, thereby resulting in a prolongation of isovolumic contraction time. Furthermore, the ability of myocardial myocytes to maintain this high LV pressure decreases, resulting in the reduction in ejection time. As LV diastolic function declines further, early dia-stolic relaxation proceeds more slowly, explaining the prolongation of isovolumic relaxation time. Consequently, the myo-cardial performance index, defined as (isovolumic relaxation time+isovolumic contraction time)/ejection time, will increase, regardless of whether the LV is suffering from impaired systolic or diastolic function. Myocardial performance index has been demonstrated to increase even in patients with severe diastolic dysfunction, in whom a short isovolumic relaxation time is seen. This is attributable to the increase in isovolumic contraction time or decrease in ejection time. The new tissue Doppler imaging M-mode method of obtaining cardiac time intervals is easy, has a high reproducibility, and obtains all of the time intervals from the same cardiac cycle. The present study demonstrates that in patients with STEMI treated with primary percutaneous coronary intervention, the combined indexes of systolic and diastolic function (isovolumic relaxation time/ejection time and myocardial performance index) provide independent prognostic information, regardless of rhythm, incremental to conventional and novel echocardiographic parameters of systolic and diastolic function.

1

The Cardiac Time Intervals measured by Tissue

Doppler Imaging M-mode: Association to

Hypertension, Left Ventricular Geometry and

Future Ischemic Cardiovascular Diseases

Tor Biering-Sørensen*†‡, MD; Rasmus Mogelvang*†, MD, PhD,

Peter Schnohr† MD, DMSc, and Jan Skov Jensen*†‡, MD, PhD, DMSc.

*Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark

†The Copenhagen City Heart Study, Bispebjerg Hospital, University of Copenhagen, Denmark

‡Institute of Clinical Medicine, Faculty of Health Sciences, University of Copenhagen, Denmark

Running title: Cardiac Time Intervals and Hypertension.

Word count: 6,845 (includes title page, abstract, text, references, tables, and figure legends)

Subject Codes: Hypertension:[13] Cerebrovascular disease/stroke; Hypertension:[15]

Hypertrophy; Hypertension:[115] Remodeling; Diagnostic testing:[31] Echocardiography;

Atherosclerosis:[150] Imaging

Correspondence:

Tor Biering-Sørensen, MD

Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark

Niels Andersensvej 65, DK-2900, Post 835, Copenhagen (Denmark),

Phone: +45 28933590 / +1 (601) 901 2765, Fax: +45 39777381

E-mail: tor.biering@gmail.com

2

Abstract

Background: We hypothesized that the cardiac time intervals reveal reduced myocardial function

in persons with hypertension and are strong predictors of future ischemic cardiovascular diseases

(ICVD) in the general population.

Methods and Results: In a large community based population study, cardiac function was

evaluated in 1,915 participants by both conventional echocardiography and by tissue Doppler

imaging (TDI). The cardiac time intervals, including the isovolumic relaxation time (IVRT),

isovolumic contraction time (IVCT), and the ejection time (ET) were obtained by TDI M-mode

through the mitral leaflet. IVCT/ET, IVRT/ET and the myocardial performance index

(MPI=(IVRT+IVCT)/ET) were calculated.

After multivariable adjustment for clinical variables the IVRT, IVRT/ET and MPI, remained

significantly impaired in persons with hypertension (n=826) compared to participants without

(n=1082). Additionally, they displayed a significant dose-response relationship, between increasing

severity of elevated blood pressure and increasing left ventricular mass index (p<0.001 for all).

Furthermore, during follow-up of median 10.7 years 435 had an ICVD (ischemic heart disease,

peripheral arterial disease or stroke). The IVRT/ET and MPI were powerful and independent

predictors of future cardiovascular disease especially in participants with known hypertension. They

provide prognostic information incremental to clinical variables from the Framingham Risk Score,

the SCORE risk chart and the ESH/ESC risk chart.

Conclusion: The cardiac time intervals identify impaired cardiac function in individuals with

hypertension, not only independent of conventional risk factors, but also in participants with a

normal conventional echocardiographic examination. The IVRT/ET and MPI are independent

predictors of future cardiovascular disease especially in participants with known hypertension.

3

Keywords: Risk-stratification; Cardiac Time Intervals; Hypertension; Ischemic Cardiovascular

Disease; Tissue Doppler Imaging.

4

Introduction

Hypertension is very common in the general population where the prevalence is reported to be

around 30-45%1. Increasing blood pressure (BP) and hypertension are strong predictors of future

ischemic cardiovascular diseases (ICVD) such as stroke, ischemic heart disease (IHD) and

peripheral artery disease (PAD)2–4

. The risk of ICVDs increases continuously with increasing BP

without any evidence of a lower threshold5. In addition, the presence of elevated BP leads to organ

damage in the heart, eyes, blood vessels, kidneys and the brain, all of which often are

asymptomatic1. Asymptomatic organ damage is an intermediate in the continuum from elevated BP

to fulminant ICVD. When asymptomatic organ damage of the heart has occurred, the progression to

fulminant ICVD is imminent explaining the increased risk of ICVD with the observation of

asymptomatic organ damage1,6–8

. Increased BP and hypertension leads to specific changes in the

heart characterized by left ventricular hypertrophy (LVH) and diastolic dysfunction1,9,10

. Including

echocardiography in the risk stratification of patients with hypertension has been demonstrated

valuable6,11,12

, however, echocardiography is not recommended as a part of the first line of the

diagnostic workup in patients with hypertension1. Nevertheless, since the risk of ICVD increases

continuously with increasing BP5 and LVH

7,8, the optimal echocardiographic marker for risk

stratification in hypertension related diseases, should be able to display changes in both severity of

elevated BP and left ventricular geometry, improve the current risk prediction models and be able to

identify early cardiac impairment prior to the development of asymptomatic organ damage of the

heart.

The cardiac time intervals are sensitive markers of cardiac dysfunction, even when it goes

unrecognized by conventional echocardiography13–15

and they have previously been demonstrated

to contain prognostic information incremental to the conventional echocardiographic parameters in

various populations16–19

. A novel method of obtaining the global cardiac time intervals has recently

5

evolved. Using this method, the cardiac time intervals are obtained by evaluating the mitral valve

(MV) movement through the cardiac cycle using a simple color tissue Doppler imaging (TDI) M-

mode analysis19–22

.

Nevertheless, it is unknown if the cardiac time intervals are able to identify miniscule cardiac

impairments in individuals with hypertension, which are unrecognized by conventional

echocardiography, and if these time intervals are affected according to BP severity and left

ventricular geometry. Furthermore, it is unknown if these cardiac time intervals can be used to

predict hypertension related diseases such as ICVD.

6

Methods

Study population: Within the Copenhagen City Heart Study, a longitudinal cohort study of

cardiovascular disease and risk factors, an echocardiographic sub study was performed22,23

. The

present study includes all participants from the fourth Copenhagen City Heart Study examination

2001-2003, who had an echocardiographic examination, including TDI, performed. We hereby

obtained the cardiac time intervals from 1,915 participants aged 20 to 93 years. Whether a

participant underwent echocardiography as part of the fourth Copenhagen City Heart Study

examination was independent of his or her health status and other risk factors.

All participants gave written informed consent, and the study was performed in accordance with the

second Helsinki Declaration and approved by the regional ethics committee.

Health Examination

All participants answered a self-administered questionnaire and underwent thorough physical

examinations. BP of all participants was measured with the London School of Hygiene

sphygmomanometer. Hypertension was defined as systolic BP of 140 mm Hg or above, diastolic

BP of 90 mm Hg or above, or use of antihypertensive medication24

. A 12-lead ECG was recorded at

rest in a supine position and coded according to the Minnesota code. Plasma cholesterol and blood

glucose values were measured on non-fasting venous blood samples25

. Diabetes mellitus was

defined as plasma glucose concentration of 11.1 mmol/L or above, use of insulin or other

antidiabetic medicine, self-reported disease, or hemoglobin A1c level of 7.0% or above26,27

.

Ischemic heart disease (IHD) was defined as a history of hospital admission for acute coronary

artery occlusion, percutaneous coronary intervention or coronary artery bypass grafting, or major

ischemic alterations on the ECG as defined by Minnesota codes 1.1 to 3.

7

Echocardiography: All echocardiograms were obtained using Vivid 5 ultrasound systems (GE

Healthcare, Horten, Norway) with a 2.5-MHz transducer by three sonographers. All participants

were examined with conventional two-dimensional echocardiography and color TDI. All

echocardiograms were stored on magneto-optical disks and an external FireWire hard drive (LaCie,

France) and analyzed offline with commercially available software (EchoPac, GE Medical, Horten,

Norway) by two investigators (TBS and RM), who were blinded to all other information.

Conventional echocardiography: Regional function was evaluated by the 16 standard segments

model, as suggested by the American Society of Echocardiography9. Left ventricular ejection

fraction (LVEF) was evaluated by 1 observer on the basis of the wall motion index score. LV

systolic dysfunction was defined as LVEF below 50%.

An M-mode still frame between the tips of the mitral leaflets and the tips of the papillary muscles

was recorded in the parasternal long-axis view. If the correct 90° angle to the long axis of the

ventricle could not be obtained, 2-dimensional images were used instead to quantify the myocardial

thickness and the dimensions of the LV in end-diastole (LVIDd) and the left atrium in end-systole

(LAD). LV mass index (LVMI) was calculated as the anatomic mass9 divided by body surface

area28

. Pulsed wave Doppler at the apical view was utilized to record mitral inflow between the tips

of the mitral leaflets. Peak velocity of early (E) and atrial (A) diastolic filling and deceleration time

of the E-wave (DT) were measured and the E/A-ratio was calculated. Left ventricular hypertrophy

(LVH) was defined as LVMI ≥ 96 g/m2 for women and ≥ 116 g/m

2 for men

9. Left ventricular

dilatation was considered present if LVIDd/height ≥ 3.3 cm/m9.

Diastolic dysfunction was defined as DT < 140 ms and E/A<50 years > 2.5, E/A50–70 years > 2, or E/A>70

years > 1.5, respectively.

8

Participants were stratified according to left ventricular geometry. Normal geometry was defined as

a relative wall thickness (RWT) ≤ 0.42 and absence of LVH, concentric remodeling was defined as

a RWT > 0.42 and absence of LVH, eccentric hypertrophy was defined as a RWT ≤ 0.42 and the

presence of LVH and concentric hypertrophy was defined as a RWT > 0.42 and the presence of

LVH9.

A normal conventional echocardiographic examination was defined as the absence of left

ventricular hypertrophy, dilatation, LVEF < 50%, and diastolic dysfunction.

Tissue Doppler imaging: Color TDI tracings were obtained with the range gate placed at the septal

and lateral mitral annulus in the 4-chamber view. The peak longitudinal early diastolic (e’) velocity

was measured and the average was calculated from the lateral and septal velocities and used to

obtain the E/e’.

By placing a 2-4 cm straight M-mode line through the septal half of the mitral leaflet in the color

TDI 4-chamber view, the cardiac time intervals were measured directly from the color diagram19,22

(Figure 1). The IVCT was defined as the time interval from the MV closure (MVC), determined by

the color shift from blue/turquoise to red at end-diastole, to the aortic valve opening (AVO)

determined by the color shift from blue to red (Figure 1). The ET was defined as the time interval

from AVO to aortic valve closing (AVC), determined by the color shift from red to blue at end

systole (Figure 1). The IVRT was defined as the time interval from the AVC to the MV opening

(MVO), determined by the color shift from red-orange to yellow (Figure 1). The method has

previously been validated 19,21,22

, and demonstrated superior to the conventional method of

obtaining the cardiac time intervals21,22

. Furthermore, we have previously demonstrated high

reproducibility of the method in this population22

. Both isovolumic time intervals were divided with

ET creating IVRT/ET and IVCT/ET, respectively, and MPI was calculated as the sum of the two

9

((IVRT+IVCT)/ET).

Follow-up and end Outcome: The primary endpoint of ICVD was the combined endpoint of being

admitted with IHD, peripheral arterial disease (PAD) or stroke. Follow-up was 100%. Follow-up

data on admission with IHD, PAD or stroke were obtained from the Danish National Board of

Health’s National Patient Registry, using ICD-10.

Statistics: Proportions were compared using χ2-test, continuous Gaussian distributed variables with

Student’s t-test and Mann-Whitney test if non-Gaussian distributed. The association between the

cardiac time intervals and hypertension was tested by univariable and multivariable regression

analyses including significant confounders detected from the baseline differences between

participants with and without hypertension (Table 1). Linearity, variance homogeneity, and the

assumption of normality were tested with plots of residuals. Trends were analyzed by linear

regression analyses and departure from linearity was assessed by simultaneous assessment of linear

and quadratic effects. Subdistribution Hazard Ratios (SHR) were calculated by competing risk Cox

proportional hazards regression analyses. The assumptions of proportional hazards in the models

were tested graphically and tested based on the Schoenfeld residuals. Predictive models, using

logistic regression models, predicting the risk of future hypertension related ICVD, for participants

followed for median 10.7 years, were constructed. Reclassification analysis using risk categories of

0% to 5%, 5% to 10%, 10% to 20% and >20% for low, intermediate, high and very high risk

categories1 were used to assess the net reclassification index (NRI) when adding the IVRT/ET, the

MPI, LVEF and diastolic dysfunction to the significant clinical predictors from the Framingham

Risk Score, the Systemic Coronary Evaluation (SCORE) risk chart and the European Society of

Hypertension/European Society of Cardiology (ESH/ESC risk chart). Non-Gaussian distributed

10

continues variables (LVMI and E/e’) were categorised into quartiles when included as confounders

in the models. A p-value ≤ 0.05 in 2-sided test was considered statistically significant. All analyses

were performed with STATA Statistics/Data analysis, SE 12.0 (StataCorp, Texas,USA).

11

Results

The Cardiac time intervals and hypertension: Participants with hypertension displayed impaired

systolic and diastolic function determined by affected systolic and diastolic cardiac time intervals

(Table 1). After multivariable adjustment for clinical variables, the IVRT, the IVCT, and the

combined indexes including information about systolic and diastolic function, the IVRT/ET, and

the MPI, remained independently impaired in participants with hypertension (Figure 2 and S1).

However, the IVCT did not remain significantly impaired in participants with hypertension when

the analysis was confined to persons with a normal conventional echocardiography (Figure S1). In

contrast, the IVRT, the IVRT/ET, and the MPI, remained significantly impaired after multivariable

adjustment when the analysis was confined to persons with a normal conventional

echocardiography (Figure 2).

The Cardiac time intervals and elevated blood pressure: To investigate if cardiac dysfunction

as assessed by cardiac time intervals was associated to severity of elevated BP our study population

was stratified into four groups according to BP: Normal, high normal, hypertension stage 1, and

hypertension stage 2 (Figure 3 and S2). The IVRT and the combined indexes including information

about systolic and diastolic function, the IVRT/ET and the MPI, increased incrementally with

increasing severity of elevated BP (Figure 3). This pattern of incremental increase with increasing

severity of elevated BP remained statistically significant after multivariable adjustment for both

clinical variables and conventional echocardiographic parameter (Figure 3). The same incremental

pattern was found when stratifying the participants according to the tertiles of the mean arterial

blood pressure (Figure 4 and S3). In order to determine if the cardiac time intervals can detect

miniscule cardiac dysfunction preceding fulminant hypertension, we confined the analysis to

participants without hypertension and found that both the IVRT, IVRT/ET and the MPI was

12

significantly higher in participants with high-normal compared to the group with normal BP

(normal vs. pre-hypertensive: IVRT: 88 ms (CI 87-90) vs. 98 ms (CI 97-100), p<0.001; IVRT/ET:

0.31 (CI 0.30-0.31) vs. 0.35 (CI 0.34-0.36), p<0.001; MPI: 0.43 (CI 0.42-0.44) vs. 0.48 (CI 0.47-

0.49), p<0.001).

The Cardiac time intervals and left ventricular geometry:

All the cardiac time intervals displayed impaired cardiac function with increasing values of the

IVRT, IVCT, IVRT/ET, IVCT/ET, MPI (β-coefficients: IVRT: 2.15 ms; IVCT: 1.23 ms;

IVRT/ET: 0.01; IVCT/ET: 0.01; MPI: 0.02) and decreasing value of the ET (β-coefficient: ET: -

1.68) for increasing LVMI (pr. 10 g/m2 increase)(p<0.001 for all)(Figure 4 and S3). This pattern

remained statistically significant for all the cardiac time intervals even after multivariable

adjustment for age, gender, BMI, eGFR, heart rate, systolic and diastolic BP, diabetes, cholesterol,

atrial fibrillation, IHD, IS, LVEF<50%, LVIDd, LAD, DT, E/A-ratio<1 and E/e’(p<0.05 for all).

All the cardiac time intervals, except IVCT, displayed the same pattern with increasing values of

the IVRT, IVRT/ET, IVCT/ET, MPI (β-coefficients: IVRT: 5.743 ms; IVRT/ET: 0.025; IVCT/ET:

0.004; MPI: 0.029) and decreasing value of the ET (β-coefficient: ET: -3.733 ms) for increasing

RWT (pr. 10% increase)(p<0.02 for all). This pattern also remained statistically significant for all

the cardiac time intervals, except for the IVRT, after multivariable adjustment for the same

variables as listed above (p<0.05 for all).

Participants were divided into four groups according to their left ventricular geometry (Figure 5 and

S4). Only the IVRT, IVRT/ET and MPI displayed impaired cardiac function in all the pathological

left ventricular geometry groups (concentric remodeling, concentric hypertrophy and eccentric

hypertrophy) compared to participants with normal geometry (p<0.001 for all)(Figure 5). After

multivariable adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic and

13

diastolic BP, diabetes, cholesterol, atrial fibrillation, IHD, IS, LVEF<50%, LVIDd, LAD, DT, E/A-

ratio<1 and E/e’, only the combined indexes including information about systolic and diastolic

function, the IVRT/ET, and the MPI, remained statistically significantly impaired in all the

pathological left ventricular geometry groups (Figure 5).

Usefulness of the cardiac time intervals to predict ICVD:

During follow-up of median 10.7 (interquartile range 7.3-11.1) years, 435 participants reached the

combined endpoint of ICVD (IHD, stroke or PAD). In univariable analysis, all the cardiac time

intervals were significant predictors of future ICVD (p<0.001)(Figure 5a). After multivariable

adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic BP, diastolic BP,

diabetes, cholesterol, smoking status, atrial fibrillation, IHD, IS, LVEF<50%, diastolic dysfunction

and LV hypertrophy, only the combined indexes including information about systolic and diastolic

function, the IVRT/ET, and the MPI, remained independent predictors of future ICVD (IVRT/ET:

SHR: 1.11 (1.00-1.23) per 0.1 increase, p=0.050; MPI: SHR: 1.11 (1.01-1.21) per 0.1 increase,

p=0.024)(Figure 5c). However, hypertension significantly modified the relationship between IVRT,

IVRT/ET, MPI and ICVD (p for interaction: p≤0.001 for all). The IVRT, IVRT/ET and MPI

primarily predicted ICVD in participants with hypertension but not in participants without

hypertension (Figure 5d). Furthermore, only the MPI remained an independent predictor of ICVD

in participants with hypertension (MPI: SHR: 1.12 (1.01-1.23) per 0.1 increase, p=0.027)(Figure

5d). In addition, reclassification analysis demonstrated, that adding IVRT/ET or MPI, to the

clinical predictors from the Framingham Risk score29

and the SCORE risk chart30

(age, gender,

cholesterol, smoking status and systolic blood pressure), yielded better predicting models with

significantly increase in the categorical net reclassification index (NRI) of 2.27% (95% CI 0.17-

4.37%, p=0.034) for IVRT/ET and 4.02% (95% CI 1.50-6.53%, p=0.002) for the MPI, respectively.

14

Additionally, reclassification analysis demonstrated, that adding IVRT/ET or MPI, to a model also

including the clinical predictors from the newer ESH/ESC risk chart1 (age, gender, smoking status,

cholesterol, diabetes, systolic BP, diastolic BP, LVH, chronic kidney disease (defined as eGFR≤60

mL/min/1,73 m2), IHD and IS), yielded better predicting models with significantly increase in the

categorical NRI of 3.41% (95% CI 0.79-6.03%, p=0.011) for IVRT/ET and 3.44% (95% CI 0.39-

6.48%, p=0.027) for the MPI, respectively. In contrast, when adding LVEF or diastolic dysfunction

to either of the aforementioned models, the predicting models did not improve (When added to the

to the clinical predictors from the Framingham Risk score29

and the SCORE risk chart30

: p for

LVEF 0.62; p for diastolic dysfunction 0.82; when added to the clinical predictors from the newer

ESH/ESC risk chart1: p for LVEF 0.90; p for diastolic dysfunction 0.12).

15

Discussion

In the largest prospective study to date, of a random sample of participants from the general

population undergoing comprehensive echocardiography including an assessment of the cardiac

time intervals by TDI M-mode, we found that: – The cardiac time intervals (IVRT, IVRT/ET, and

MPI) display impaired cardiac function in individuals with hypertension, not only independent of

conventional risk factors, but also in participants with a normal conventional echocardiographic

examination. – A significant dose-response relationship, between increasing severity of BP and

incremental impairment in cardiac function determined by the cardiac time intervals (IVRT,

IVRT/ET, and MPI) exists. – The combined indexes encompassing information on both systolic

and diastolic function (the IVRT/ET and the MPI), independently display cardiac dysfunction in all

types of pathological left ventricular geometry. – The IVRT/ET and the MPI are independent

predictors of future ICVD, even after adjustment for conventional echocardiographic parameters.

However, they primary provide prognostic information regarding the risk of future ICVD in

participants with hypertension. – The IVRT/ET and the MPI improve the predictive models for

future ICVD when added to the Framingham Risk Score, the SCORE risk chart and the ESH/ESC

risk chart.

The Cardiac time intervals, hypertension and left ventricular geometry: Cardiac function was

significantly impaired in participants with hypertension compared to participants with normal blood

pressure, determined by affected systolic and diastolic cardiac time intervals (Table 1). However,

after multivariable adjustment, and when confining the analysis to participants with a normal

conventional echocardiographic examination, only IVRT, IVRT/ET, and MPI remained

significantly impaired in participants with hypertension (Figure 2 and S1). These results were

expected, since it is well known that patients with hypertension primarily suffer from an impaired

16

diastolic function and not an impaired systolic function31,32

, why only the time intervals containing

information about the diastolic function remained impaired after multivariable adjustment.

However, the pivotal finding in Figure 2 was that the cardiac time intervals were capable of

identifying subtle impairments in the cardiac function in participants with hypertension, which were

unnoticed by conventional echocardiography (Figure 2). Therefore, the cardiac time intervals,

containing information about diastolic function, seem capable of identifying participants in the risk

of hypertension related asymptomatic organ damage of the heart, despite of a normal conventional

echocardiography.

In accordance with our results, a previous small scale study including 62 hypertensive patients and

15 controls, also found IVRT and MPI to be the only cardiac time intervals demonstrating impaired

cardiac function in patients with hypertension, despite of a normal systolic function33

.

Unfortunately, Cacciapuoti and colleagues did not evaluate the diagnostic utility of IVRT/ET in

their study33

, however if they had , it would most likely also be impaired in the hypertensive group.

We found a significant linear dose-response relationship, between increasing severity of elevated

blood pressure and incremental impairment in cardiac function determined by the age and gender

adjusted cardiac time intervals (IVRT, IVRT/ET and MPI)(Figure 3). Even after multivariable

adjustment for all other clinical and conventional echocardiographic parameters this pattern of

incremental increase in cardiac dysfunction determined by IVRT, IVRT/ET and MPI, with

increasing severity of elevated blood pressure, remained statistically significant.

Likewise, a previous study demonstrated that MPI increased with increasing severity of

hypertension according to left ventricular geometry34

. Accordingly, we found that all the cardiac

time intervals displayed impaired cardiac function with increasing values of the LVMI (Figure 4

and S3) even after multivariable adjustment. In addition, the IVRT/ET and the MPI displayed

cardiac dysfunction after multivariable adjustment in all pathological left ventricular geometry

17

groups (Figure 5). These findings of incremental impairment of the cardiac time intervals

(especially the IVRT/ET and MPI) with increasing BP and LVMI (Figure 4 and S3), and in all

types of pathological left ventricular geometry, are very important, because a potential future

echocardiographic parameter for risk stratification in hypertension, should display incremental

impairment in the measure with increasing BP, LVMI and in all types of pathological left

ventricular geometry, since the risk of ICVD increases continuously with increasing BP5 and

LVH7,8

.

In addition, another central observation in our study was that IVRT and MPI were capable in

detecting miniscule cardiac dysfunction in participants with high normal BP compared to

participants with normal BP, when confining our analysis to participants without hypertension and

after adjustment for clinical variables. This is very important because it has previously been

demonstrated that persons with high normal BP have increased risk of developing future

hypertension35

and increased risk of ICVD36

. Thus, IVRT and MPI reveal impaired cardiac function

in persons with high normal BP, why these time intervals may aid in the early identification of

persons in high risk of future fulminant hypertension and ICVD.

The cardiac time intervals and the risk for ICVD:

It is well known that hypertension is associated with increased risk of ICVD3, however the

usefulness of the cardiac time intervals to predict future ICVD in the general population and in

participants with hypertension is unknown. We found that only the combined indexes containing

information on both systolic and diastolic performance (IVRT/ET and MPI) were independent

predictors of future ICVD, even after adjustment for hypertension status, LVH and systolic and

diastolic BP (Figure 5c). Consequently, even though the IVRT/ET and MPI both are independently

associated with BP severity and LVMI, they provide independent prognostic information to both

18

BP and LVH. The superiority of the IVRT/ET and the MPI to predict future ICVD may be due to

the fact that they detect very miniscule impairment in the myocardial function, irrespective if the

myocardium suffers from an ailing systolic or diastolic function19

. Therefore, in the general

population, the IVRT/ET and MPI identifies subtle impairments in the cardiac function (both

systolic and diastolic), that may be unnoticed by conventional echocardiography13,14

(Figure 2), and

can improve risk stratification strategies evaluating future risk of fulminant hypertension related

ICVD incremental to the traditional predictors. In accordance with this, adding IVRT/ET or MPI,

to the clinical predictors from the Framingham Risk Score29

, the SCORE risk chart30

and the

ESH/ESC risk chart1 yielded better predicting models with significant increases in the categorical

NRIs. However, we found that hypertension significantly modified the relationship between all the

cardiac time intervals containing information about diastolic function (IVRT, IVRT/ET and MPI)

and the risk of future ICVD. These cardiac time intervals primarily predicted hypertension related

ICVD in participants with hypertension but not in participants without hypertension (Figure 5d).

Similarly, a previous study demonstrated that the presence of diastolic dysfunction, determined by a

short deceleration time, only predicted all-cause mortality in hypertensive heart failure patients and

not in heart failure patients without hypertension37

. Therefore, in participants suffering from

hypertension, who are in particularly high risk of future ICVD3, the IVRT/ET and MPI may be

useful cardiac measures for risk stratification. They are simple to obtain, have high

reproducible19,21,22

and encompasses information on both BP severity, left ventricular geometry and

risk of future ICVD.

Limitations

The inhabitants of Denmark and the study population are primarily Caucasian, which limits the

generalizability of our findings to other general populations with another composition.

19

Participants with short-lived or prolonged periods of hypertension were treated similarly in the

analyses.

Conclusion

The cardiac time intervals (especially the IVRT, IVRT/ET, and MPI) displayed a significant dose-

response relationship, with increasing severity of elevated BP and increasing left ventricular mass

index. Furthermore, they identified impaired cardiac function in participants with hypertension, not

only independent of conventional risk factors, but also in participants with a normal conventional

echocardiographic examination. Additionally, in the general population, the IVRT/ET and MPI, are

powerful and independent predictors of future hypertension related ICVD especially in participants

with known hypertension. They provide prognostic information incremental to clinical variables

from the Framingham Risk Score, the SCORE risk chart and the ESH/ESC risk chart.

Funding Sources

This study was financially supported by the Faculty of Health and Medical Sciences, University of

Copenhagen, Denmark. The sponsor had no role in the study design, data collection, data analysis,

data interpretation, or writing of the manuscript.

Conflict of interest

None declared

20

References

1. Mancia G, Fagard R, Narkiewicz K, Redon J, Zanchetti A, Böhm M, Christiaens T, Cifkova

R, De Backer G, Dominiczak A, Galderisi M, Grobbee DE, Jaarsma T, Kirchhof P, Kjeldsen

SE, Laurent S, Manolis AJ, Nilsson PM, Ruilope LM, Schmieder RE, Sirnes PA, Sleight P,

Viigimaa M, Waeber B, Zannad F, Redon J, Dominiczak A, Narkiewicz K, Nilsson PM,

Burnier M, Viigimaa M, Ambrosioni E, Caufield M, Coca A, Olsen MH, Schmieder RE,

Tsioufis C, van de Borne P, Zamorano JL, Achenbach S, Baumgartner H, Bax JJ, Bueno H,

Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J,

Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA,

Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Clement DL, Coca A, Gillebert

TC, Tendera M, Rosei EA, Ambrosioni E, Anker SD, Bauersachs J, Hitij JB, Caulfield M, De

Buyzere M, De Geest S, Derumeaux GA, Erdine S, Farsang C, Funck-Brentano C, Gerc V,

Germano G, Gielen S, Haller H, Hoes AW, Jordan J, Kahan T, Komajda M, Lovic D,

Mahrholdt H, Olsen MH, Ostergren J, Parati G, Perk J, Polonia J, Popescu BA, Reiner Z,

Rydén L, et al. 2013 ESH/ESC guidelines for the management of arterial hypertension: the

Task Force for the Management of Arterial Hypertension of the European Society of

Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J.

2013;34:2159–2219.

2. Lawes CMM, Rodgers A, Bennett DA, Parag V, Suh I, Ueshima H, MacMahon S, Asia

Pacific Cohort Studies Collaboration. Blood pressure and cardiovascular disease in the Asia

Pacific region. J Hypertens. 2003;21:707–716.

3. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A,

Stamler J. Blood pressure, stroke, and coronary heart disease. Part 1, Prolonged differences in

21

blood pressure: prospective observational studies corrected for the regression dilution bias.

Lancet. 1990;335:765–774.

4. Fowkes FGR, Rudan D, Rudan I, Aboyans V, Denenberg JO, McDermott MM, Norman PE,

Sampson UKA, Williams LJ, Mensah GA, Criqui MH. Comparison of global estimates of

prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review

and analysis. Lancet. 2013;382:1329–1340.

5. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R, Prospective Studies Collaboration.

Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of

individual data for one million adults in 61 prospective studies. Lancet. 2002;360:1903–1913.

6. Sehestedt T, Jeppesen J, Hansen TW, Wachtell K, Ibsen H, Torp-Pedersen C, Torp-Petersen

C, Hildebrandt P, Olsen MH. Risk prediction is improved by adding markers of subclinical

organ damage to SCORE. Eur Heart J. 2010;31:883–891.

7. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic implications of

echocardiographically determined left ventricular mass in the Framingham Heart Study. N

Engl J Med. 1990;322:1561–1566.

8. Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventricular mass

and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern

Med. 1991;114:345–352.

9. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH,

Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MSJ, Stewart WJ.

Recommendations for chamber quantification: a report from the American Society of

Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification

22

Writing Group, developed in conjunction with the European Association of Echocardiography,

a branch of the European Society of Cardiology. J Am Soc Echocardiogr Off Publ Am Soc

Echocardiogr. 2005;18:1440–1463.

10. Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth OA, Waggoner AD,

Flachskampf FA, Pellikka PA, Evangelista A. Recommendations for the evaluation of left

ventricular diastolic function by echocardiography. J Am Soc Echocardiogr Off Publ Am Soc

Echocardiogr. 2009;22:107–133.

11. Cuspidi C, Ambrosioni E, Mancia G, Pessina AC, Trimarco B, Zanchetti A, APROS

Investigators. Role of echocardiography and carotid ultrasonography in stratifying risk in

patients with essential hypertension: the Assessment of Prognostic Risk Observational Survey.

J Hypertens. 2002;20:1307–1314.

12. Cuspidi C, Meani S, Valerio C, Fusi V, Sala C, Zanchetti A. Left ventricular hypertrophy and

cardiovascular risk stratification: impact and cost-effectiveness of echocardiography in

recently diagnosed essential hypertensives. J Hypertens. 2006;24:1671–1677.

13. Ruggiero A, De Rosa G, Rizzo D, Leo A, Maurizi P, De Nisco A, Vendittelli F, Zuppi C,

Mordente A, Riccardi R. Myocardial performance index and biochemical markers for early

detection of doxorubicin-induced cardiotoxicity in children with acute lymphoblastic

leukaemia. Int J Clin Oncol. 2013;18:927–933.

14. Varol E, Akcay S, Ozaydin M, Ozturk O, Cerci SS, Sahin U. Influence of obstructive sleep

apnea on left ventricular mass and global function: sleep apnea and myocardial performance

index. Heart Vessels. 2010;25:400–404.

23

15. Kaya MG, Simsek Z, Sarli B, Buyukoglan H. Myocardial performance index for detection of

subclinical abnormalities in patients with sarcoidosis. J Thorac Dis. 2014;6:429–437.

16. Arnlöv J, Ingelsson E, Risérus U, Andrén B, Lind L. Myocardial performance index, a

Doppler-derived index of global left ventricular function, predicts congestive heart failure in

elderly men. Eur Heart J. 2004;25:2220–2225.

17. Arnlöv J, Lind L, Andrén B, Risérus U, Berglund L, Lithell H. A Doppler-derived index of

combined left ventricular systolic and diastolic function is an independent predictor of

cardiovascular mortality in elderly men. Am Heart J. 2005;149:902–907.

18. Møller JE, Egstrup K, Køber L, Poulsen SH, Nyvad O, Torp-Pedersen C. Prognostic

importance of systolic and diastolic function after acute myocardial infarction. Am Heart J.

2003;145:147–153.

19. Biering-Sørensen T, Mogelvang R, Søgaard P, Pedersen SH, Galatius S, Jørgensen PG, Jensen

JS. Prognostic Value of Cardiac Time Intervals by Tissue Doppler Imaging M-Mode in

Patients With Acute ST-Segment-Elevation Myocardial Infarction Treated With Primary

Percutaneous Coronary Intervention. Circ Cardiovasc Imaging. 2013;6:457–465.

20. Voigt J-U, Lindenmeier G, Exner B, Regenfus M, Werner D, Reulbach U, Nixdorff U,

Flachskampf FA, Daniel WG. Incidence and characteristics of segmental postsystolic

longitudinal shortening in normal, acutely ischemic, and scarred myocardium. J Am Soc

Echocardiogr Off Publ Am Soc Echocardiogr. 2003;16:415–423.

21. Kjaergaard J, Hassager C, Oh JK, Kristensen JH, Berning J, Sogaard P. Measurement of

cardiac time intervals by Doppler tissue M-mode imaging of the anterior mitral leaflet. J Am

Soc Echocardiogr Off Publ Am Soc Echocardiogr. 2005;18:1058–1065.

24

22. Biering-Sørensen T, Mogelvang R, Pedersen S, Schnohr P, Sogaard P, Jensen JS. Usefulness

of the myocardial performance index determined by tissue Doppler imaging m-mode for

predicting mortality in the general population. Am J Cardiol. 2011;107:478–483.

23. Mogelvang R, Sogaard P, Pedersen SA, Olsen NT, Marott JL, Schnohr P, Goetze JP, Jensen

JS. Cardiac dysfunction assessed by echocardiographic tissue Doppler imaging is an

independent predictor of mortality in the general population. Circulation. 2009;119:2679–

2685.

24. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW,

Materson BJ, Oparil S, Wright JT Jr, Roccella EJ, Joint National Committee on Prevention,

Detection, Evaluation, and Treatment of High Blood Pressure. National Heart, Lung, and

Blood Institute, National High Blood Pressure Education Program Coordinating Committee.

Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and

Treatment of High Blood Pressure. Hypertension. 2003;42:1206–1252.

25. Langsted A, Freiberg JJ, Nordestgaard BG. Fasting and nonfasting lipid levels: influence of

normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction.

Circulation. 2008;118:2047–2056.

26. Peters AL, Davidson MB, Schriger DL, Hasselblad V. A clinical approach for the diagnosis of

diabetes mellitus: an analysis using glycosylated hemoglobin levels. Meta-analysis Research

Group on the Diagnosis of Diabetes Using Glycated Hemoglobin Levels. JAMA J Am Med

Assoc. 1996;276:1246–1252.

25

27. Rowley KG, Daniel M, O’Dea K. Screening for diabetes in Indigenous populations using

glycated haemoglobin: sensitivity, specificity, post-test likelihood and risk of disease. Diabet

Med J Br Diabet Assoc. 2005;22:833–839.

28. Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and

weight be known. 1916. Nutr Burbank Los Angel Cty Calif. 1989;5:303–311; discussion 312–

313.

29. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of

coronary heart disease using risk factor categories. Circulation. 1998;97:1837–1847.

30. Conroy RM, Pyörälä K, Fitzgerald AP, Sans S, Menotti A, De Backer G, De Bacquer D,

Ducimetière P, Jousilahti P, Keil U, Njølstad I, Oganov RG, Thomsen T, Tunstall-Pedoe H,

Tverdal A, Wedel H, Whincup P, Wilhelmsen L, Graham IM, SCORE project group.

Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur

Heart J. 2003;24:987–1003.

31. Lamb HJ, Beyerbacht HP, van der Laarse A, Stoel BC, Doornbos J, van der Wall EE, de Roos

A. Diastolic dysfunction in hypertensive heart disease is associated with altered myocardial

metabolism. Circulation. 1999;99:2261–2267.

32. Nakashima Y, Nii T, Ikeda M, Arakawa K. Role of left ventricular regional nonuniformity in

hypertensive diastolic dysfunction. J Am Coll Cardiol. 1993;22:790–795.

33. Cacciapuoti F, Scognamiglio A, Paoli VD, Romano C, Cacciapuoti F. Left Atrial Volume

Index as Indicator of Left Ventricular Diastolic Dysfunction: Comparation between Left Atrial

Volume Index and Tissue Myocardial Performance Index. J Cardiovasc Ultrasound.

2012;20:25–29.

26

34. Akintunde AA, Akinwusi PO, Opadijo GO. Relationship between Tei index of myocardial

performance and left ventricular geometry in Nigerians with systemic hypertension.

Cardiovasc J Afr. 2011;22:124–127.

35. Vasan RS, Larson MG, Leip EP, Kannel WB, Levy D. Assessment of frequency of

progression to hypertension in non-hypertensive participants in the Framingham Heart Study:

a cohort study. Lancet. 2001;358:1682–1686.

36. Vasan RS, Larson MG, Leip EP, Evans JC, O’Donnell CJ, Kannel WB, Levy D. Impact of

high-normal blood pressure on the risk of cardiovascular disease. N Engl J Med.

2001;345:1291–1297.

37. Andersson C, Gislason GH, Weeke P, Kjaergaard J, Hassager C, Akkan D, Møller JE, Køber

L, Torp-Pedersen C, EchoCardiography and Heart Outcome Study (ECHOS) investigators.

The prognostic importance of a history of hypertension in patients with symptomatic heart

failure is substantially worsened by a short mitral inflow deceleration time. BMC Cardiovasc

Disord. 2012;12:30.

27

Figure legends.

Figure 1: The cardiac time intervals assessed by a color tissue Doppler imaging (TDI) M-

mode line through the mitral leaflet.

Left: Four-chamber gray-scale (bottom) and color TDI (top) views in end-systole displaying the

position of the M-mode line used for measuring the cardiac time intervals. Right: Color diagram of

the TDI M-mode line through the mitral leaflet.

MV=Mitral Valve; MVC=MV Closing; AVO=Aortic Valve Opening; AVC=Aortic Valve Closure;

MVO=MV Opening.

Figure 2: Systolic and diastolic function determined by the cardiac time intervals in

participants with and without hypertension after multivariable adjustment.

Multivariable adjusted means are depicted for the cardiac time intervals in participants with and

without hypertension for both the entire study cohort and the subgroup with a normal conventional

echocardiographic examination. Multivariable adjustment designates adjustment for age, gender,

body mass index, estimated glomerular filtration rate, heart rate, diabetes, cholesterol, atrial

fibrillation, ischemic heart disease and previous ischemic stroke. Normal echo designates the

absence of left ventricular hypertrophy, dilatation and ejection fraction<50%, and severe diastolic

dysfunction. Bars indicate standard error. IVRT=Isovolumic Relaxation Time; ET=Ejection Time.

Figure 3: Systolic and diastolic function determined by the cardiac time intervals according to

elevated blood pressure severity.

Depicting the adjusted means of the cardiac time intervals in participants stratified according to

blood pressure, after adjustment for age and gender. The study population was divided into four

groups according to blood pressure: normal (< 120/80 mmHg), high normal (between 120/80 and

28

140/90 mmHg), stage 1 hypertension (between 140/90 and 160/100 mmHg), and stage 2

hypertension (≥160/100 mmHg). The pattern of incremental increase of the IVRT, IVRT/ET and

MPI with increasing severity of elevated BP, remained statistically significant after multivariable

adjustment for age, gender, BMI, eGFR, heart rate, diabetes, cholesterol, atrial fibrillation, IHD, IS,

LVEF<50%, LVIDd, LAD, LVMI, DT, E/A-ratio<1 and E/e’. Bars indicate standard error.

IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI=Myocardial Performance Index;

LVEF=Left Ventricular Ejection Fraction; LVMI=Left Ventricular Mass Index; RWT=Relative

Wall Thickness; LVIDd= left ventricular dimension in end-diastole; LAD=Left atrium diameter in

end-systole; E=peak transmitral early diastolic inflow velocity; A=peak transmitral late diastolic

inflow velocity; DT=deceleration time of early diastolic inflow.

Figure 4: The association between the cardiac time intervals and LVMI and MAP

The association between the combined cardiac time intervals (IVRT/ET and MPI) and the left

ventricular mass index and mean arterial blood pressure. Participants were stratified into tertiles of

the LVMI and MAP. Bars indicate standard error. LVMI=Left Ventricular Mass Index;

MAP=Mean arterial blood pressure; IVRT=Isovolumic Relaxation Time; ET=Ejection Time;

MPI=Myocardial Performance Index.

Figure 5: The cardiac time intervals in participants with normal left ventricular geometry

(n=1122), concentric remodeling (n=259), eccentric hypertrophy (n=177) and concentric

hypertrophy (n=103)

Depicting the cardiac time intervals in participants with normal left ventricular geometry and

pathological left ventricular geometry. Displays mean and standard errors. † indicates p≤0.01 and ‡

indicates p≤0.05 after adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic and

29

diastolic BP, diabetes, cholesterol, atrial fibrillation, IHD, previous ischemic stroke, LVEF<50%,

LVIDd, LA diameter, DT, E/A-ratio<1 and E/e’, when comparing the participants with normal left

ventricular geometry to participants with pathological left ventricular geometry. LVIDd= left

ventricular dimension in end-diastole; LAD=Left atrium diameter in end-systole; E=peak

transmitral early diastolic inflow velocity; A=peak transmitral late diastolic inflow velocity;

DT=deceleration time of early diastolic inflow; IVRT=Isovolumic Relaxation Time; ET=Ejection

Time; MPI=Myocardial Performance Index.

Figure 6: The cardiac time intervals as predictors of future ischemic cardiovascular disease.

Depicting the subdistribution hazard ratios (SHR) obtained from univariable analysis (Figure 5a),

adjusted for age, gender, systolic and diastolic blood pressure (Figure 5b), and multivariable

(Figure 5c) competing risk Cox proportional hazards regression models describing the cardiac time

intervals as predictors of future hypertension related ICVD. The IVRT, IVRT/ET and the MPI are

also stratified according to hypertension status (Figure 5d). Multivariable adjustment indicates

adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic blood pressure, diastolic

blood pressure, diabetes, cholesterol, smoking status, atrial fibrillation, IHD, previous ischemic

stroke, LVEF<50%, diastolic dysfunction and LV hypertrophy. Depicting the SHR and the 95%

confidence intervals. BMI=Body Mass Index; eGFR= estimated Glomerular Filtration Rate;

LVEF=Left Ventricular Ejection Fraction; IVCT=Isovolumic Contraction Time; IVRT=Isovolumic

Relaxation Time; ET=Ejection Time; MPI=Myocardial Performance Index.

30

Table 1. Population characteristics

Non-hypertensive

(n=1082)

Hypertension

(n=826)

p-value

Age (years)

51±15 68±11 <0.001

Male gender

44% 41% 0.20

Systolic Blood Pressure (mmHg) 120±12 156±18 <0.001

Diastolic Blood Pressure (mmHg) 73±9 85±12 <0.001

Heart Rate (beats per minute) 65±11 70±12 <0.001

Smoking status

- Never

- Previous

- Current

35%

31%

34%

32%

36%

32%

0.10

Diabetes 7% 15% <0.001

Cholesterol 5.3±1.1 5.8±1.1 <0.001

Previous ischemic heart disease

3% 10% <0.001

Previous ischemic stroke

0.6% 3.5% <0.001

BMI (kg/m2) 24.5±3.4 26.6±4.2 <0.001

eGFR (mL/min/1,73 m2) 79.2±15.7 72.4±16.1 <0.001

Atrial fibrillation/flutter 1% 3% 0.007

LVEF < 50% 0.5% 1.6% 0.012

LV Dilatation 4% 6% 0.037

LV Hypertrophy 9% 29% <0.001

LV Diastolic dysfunction 0.5% 2.0% 0.003

Cardiac time intervals:

IVRT (ms) 94±21 110±24 <0.001

IVCT (ms) 36±13 39±15 <0.001

ET (ms) 286±24 281±30 <0.001

IVRT/ET 0.33±0.08 0.40±0.11 <0.001

IVCT/ET 0.13±0.05 0.14±0.06 <0.001

MPI 0.46±0.11 0.54±0.15 <0.001

31

BMI=Body Mass Index; eGFR=Estimated Glomerular Filtration Rate; LVEF=Left Ventricular

Ejection Fraction; LV=Left Ventricle; IVCT=Isovolumic Contraction Time; IVRT=Isovolumic

Relaxation Time; ET=Ejection Time; MPI=Myocardial Performance Index.

32

Figure 1

33

Figure 2

34

Figure 3

35

Figure 4

36

Figure 5

37

Figure 6

38

SUPPLEMENTAL MATERIAL

Figure Legends

Figure S1: Cardiac function determined by the cardiac time intervals in participants without

and with hypertension.

Multivariable adjusted means are depicted for the cardiac time intervals in participants with and

without hypertension for both the entire study cohort and the subgroup with a normal conventional

echocardiographic examination. Multivariable adjustment designates adjustment for age, gender,

body mass index, estimated glomerular filtration rate, heart rate, diabetes, cholesterol, atrial

fibrillation, ischemic heart disease and previous ischemic stroke. Normal echo designates the

absence of left ventricular hypertrophy, dilatation and ejection fraction<50%, and mild or severe

diastolic dysfunction. Bars indicate standard error. IVCT =Isovolumic Contraction Time;

ET=Ejection Time.

Figure S2: Systolic function determined by the cardiac time intervals according to elevated

blood pressure severity.

Depicting the adjusted means of the cardiac time intervals in participants stratified according to

blood pressure, after adjustment for age and gender. The study population was divided into four

groups according to blood pressure: normal (< 120/80 mmHg), high normal (between 120/80 and

140/90 mmHg), stage 1 hypertension (between 140/90 and 160/100 mmHg), and stage 2

hypertension (≥160/100 mmHg). The pattern of incremental increase of the IVRT, IVRT/ET and

MPI with increasing severity of elevated BP, remained statistically significant after multivariable

adjustment for age, gender, BMI, eGFR, heart rate, diabetes, cholesterol, atrial fibrillation, IHD, IS,

LVEF<50%, LVIDd, LAD, LVMI, DT, E/A-ratio<1 and E/e’. Bars indicate standard error.

39

IVCT=Isovolumic Contraction Time; ET=Ejection Time; LVEF=Left Ventricular Ejection

Fraction; LVMI=Left Ventricular Mass Index; RWT=Relative Wall Thickness; LVIDd= left

ventricular dimension in end-diastole; LAD=Left atrium diameter in end-systole; E=peak

transmitral early diastolic inflow velocity; A=peak transmitral late diastolic inflow velocity;

DT=deceleration time of early diastolic inflow.

Figure S3. The association between the cardiac time intervals and LVMI and MAP

The association between the cardiac time intervals (IVRT, IVCT, ET and ICRT/ET) and the left

ventricular mass index and mean arterial blood pressure. Participants were stratified into tertiles of

the LVMI and MAP. Bars indicate standard error. LVMI=Left Ventricular Mass Index;

MAP=Mean arterial blood pressure; IVRT=Isovolumic Relaxation Time; IVCT=Isovolumic

Contraction Time; ET=Ejection Time; MPI=Myocardial Performance Index.

Figure S4: The cardiac time intervals in participants with normal left ventricular geometry

(n=1122), concentric remodeling (n=259), eccentric hypertrophy (n=177) and concentric

hypertrophy (n=103)

Depicting the cardiac time intervals in participants with normal left ventricular geometry and

pathological left ventricular geometry. Displays mean and standard errors. * indicates significant

difference (p<0.01) when comparing the participants with normal left ventricular geometry to

participants with pathological left ventricular geometry. † indicates p≤0.01 and ‡ indicates p≤0.05

after adjustment for age, gender, BMI, eGFR, heart rate, hypertension, systolic and diastolic BP,

diabetes, cholesterol, atrial fibrillation, IHD, previous ischemic stroke, LVEF<50%, LVIDd, LA

diameter, DT, E/A-ratio<1 and E/e’, when comparing the participants with normal left ventricular

geometry to participants with pathological left ventricular geometry. LVIDd= left ventricular

40

dimension in end-diastole; LAD=Left atrium diameter in end-systole; E=peak transmitral early

diastolic inflow velocity; A=peak transmitral late diastolic inflow velocity; DT=deceleration time of

early diastolic inflow; IVCT=Isovolumic Contraction Time; ET=Ejection Time.

41

Figure S1

42

Figure S2

43

Figure S3

44

Figure S4

1

Prognostic value of cardiac time intervals

measured by tissue Doppler imaging

M-mode in the General Population

Tor Biering-Sørensen*‡†, MD; Rasmus Mogelvang*‡, MD, PhD

and Jan Skov Jensen*‡†, MD, PhD, DMSc.

* Department of Cardiology, Gentofte Hospital, University of Copenhagen, Denmark

‡ The Copenhagen City Heart Study, Bispebjerg Hospital, University of Copenhagen, Denmark

† Institute of Clinical Medicine, Faculty of Health Sciences, University of Copenhagen, Denmark

Running title: Cardiac Time Intervals predicts outcome in the general population.

Word count: 6,974 (includes title page, abstract, text, references, tables, and figure legends)

Correspondence:

Tor Biering-Sørensen, MD

Department of Cardiology, Gentofte University Hospital, Copenhagen, Denmark

Niels Andersensvej 65, DK-2900, Post 835, Copenhagen (Denmark),

Phone: +45 28933590, Fax: +45 39777381

E-mail: tor.biering@gmail.com

2

Abstract

Purpose: Tissue Doppler imaging (TDI) M-mode through the mitral leaflet is an easy, fast, and

precise method to estimate the cardiac time intervals. The aim was to evaluate the usability of the

cardiac time intervals in predicting major cardiovascular events (MACE) in the general population.

Methods: In a large community based population study, cardiac function was evaluated in 1,915

participants by both conventional echocardiography and by TDI. The cardiac time intervals,

including the isovolumic relaxation time (IVRT), isovolumic contraction time (IVCT), and ejection

time (ET) were obtained by TDI M-mode through the mitral leaflet. IVCT/ET, IVRT/ET and the

myocardial performance index (MPI=(IVRT+IVCT)/ET) were calculated.

Results: During follow-up (median 10.8 years), 383 (20%) participants reached the combined

endpoint MACE (ischemic heart disease, heart failure or cardiac death). After multivariable

adjustment for clinical predictors and conventional echocardiography, only the combined indexes,

including information on both the systolic and diastolic performance (IVRT/ET and MPI), remained

significant prognosticators (p<0.05 for both). Adding IVRT/ET or MPI to a model already

including all other echocardiographic parameters resulted in a significant increase in the Harrell’s c-

statistics (p<0.05 for both). Finally, when adding IVRT/ET or MPI to the clinical predictors

improved reclassification significantly (p<0.05 for both).

Conclusion: In the general population, the combined cardiac time intervals which include

information on both the systolic and diastolic function in one index (IVRT/ET and MPI) are not

only powerful and independent predictors of future MACE, but provide additional prognostic

information to clinical and conventional echocardiographic measures of systolic and diastolic

function.

3

Introduction

Preservation of normal cardiac time intervals is intimately related to normal cardiac physiology,

hemodynamics, mechanics and biochemistry. In the ailing myocardium, the cardiac time intervals

will change during disease progression1–5

. As left ventricular (LV) systolic function deteriorates, the

time it takes for myocardial myocytes to achieve an LV pressure equal to that of aorta increases,

resulting in a prolongation of isovolumic contraction time (IVCT)5. Furthermore, the ability of

myocardial myocytes to maintain the LV pressure decreases, resulting in reduction in the ejection

time (ET)5. As LV diastolic function declines further, early diastolic relaxation proceeds more

slowly, explaining the prolongation of isovolumic relaxation time (IVRT). Consequently, the

IVRT/ET and the myocardial performance index (MPI), defined as (IVRT + IVCT)/ET, will detect

cardiac dysfunction with an increase, irrespective of whether the LV is suffering from impaired

systolic or diastolic function6–8

. The MPI increases even in patients with severe diastolic

dysfunction with a restrictive filling pattern, which is seen as a short IVRT time9. This is

attributable to the increase in IVCT or decrease in ET, illustrating the subtle impaired systolic

function identified by novel echocardiographic parameters which is found in patients with severe

diastolic dysfunction, even though they display a preserved LV ejection fraction10,11

. Therefore, the

cardiac time intervals, especially the combined indexes containing information on both systolic and

diastolic performance (the IVRT/ET and the MPI), may be useful to identify subtle impairments in

the cardiac function (both systolic and diastolic) in the general population, which are unnoticed by

conventional echocardiography12,13

.These echocardiographic parameters may thus identify patients

in high risk of future fulminant cardiovascular disease.

The novel method of obtaining the global cardiac time intervals, by using tissue Doppler imaging

(TDI) M-mode through the mitral leaflet, is easy, has a high reproducibility, and obtains all of the

time intervals from the same cardiac cycle1,6,14

.

4

We hypothesised that the cardiac time intervals, especially the combined indexes containing

information on both systolic and diastolic function (the IVRT/ET and the MPI), are strong

predictors of future cardiovascular disease and cardiovascular mortality in the general population.

Furthermore, we hypothesised that the combined indexes provides prognostic information above

and beyond clinical predictors and conventional echocardiographic measures of systolic and

diastolic function.

5

Methods

Study population: The Copenhagen City Heart Study is a longitudinal cohort study of

cardiovascular disease and risk factor1,15

. This echocardiographic sub study includes all participants

from the fourth Copenhagen City Heart Study examination round in 2001 through 2003, who had

an echocardiographic examination, including TDI, performed. The cardiac time intervals were

obtained in a total of 1915 participants aged 20 to 93 years. Whether a participant underwent

echocardiography as part of the fourth Copenhagen City Heart Study examination was independent

of his or her health status and other risk factors.

All participants gave written informed consent, and the study was performed in accordance with the

second Helsinki Declaration and approved by the regional ethics committee.

Health Examination

All participants underwent physical examinations as well as a self-administered questionnaire.

Blood pressure was measured with the London School of Hygiene sphygmomanometer. Plasma

cholesterol and blood glucose values were measured on non-fasting venous blood samples16

. A 12-

lead ECG was recorded at rest in a supine position and coded according to the Minnesota code.

Diabetes mellitus was defined as plasma glucose concentration of 11.1 mmol/L or above, use of

insulin or other antidiabetic medicine, self-reported disease, or hemoglobin A1c level of 7.0% or

above17

. Hypertension was defined as systolic blood pressure of 140 mm Hg or above, diastolic

blood pressure of 90 mm Hg or above, or use of antihypertensive medication18

. Ischemic heart

disease (IHD) was defined as a history of hospital admission for acute coronary artery occlusion,

percutaneous coronary intervention or coronary artery bypass grafting, or major

ischemic alterations on the ECG as defined by Minnesota codes 1.1 to 3.

6

Echocardiography: Echocardiography was performed using Vivid 5 ultrasound systems (GE

Healthcare, Horten, Norway) with a 2.5-MHz transducer by three experienced sonographers. All

participants were examined with conventional two-dimensional echocardiography and color TDI.

All echocardiograms were stored on magneto-optical disks and an external FireWire hard drive

(LaCie, France) and analyzed offline with commercially available software (EchoPac, GE Medical,

Horten, Norway) by two investigators (TBS and RM), who were blinded to other information.

Conventional echocardiography: The 16 standard segments model, as suggested by the American

Society of Echocardiography was used to evaluate regional function19

. Evaluation of LV ejection

fraction (LVEF) was made by 1 observer on the basis of the wall motion index score. LV systolic

dysfunction was defined as LVEF below 50%.

An M-mode still frame between the tips of the mitral leaflets and the tips of the papillary muscles

was recorded in the parasternal long-axis view. If the correct 90° angle to the long axis of the

ventricle could not be obtained, 2-dimensional images were used instead to quantify the myocardial

thickness and the dimensions of the LV and the left atrium. LV mass index was calculated as the

anatomic mass19

divided by body surface area. The mitral inflow velocity curves were measured

using pulsed-wave Doppler in the apical position. Peak velocity of early (E) and atrial (A) diastolic

filling and deceleration time of the E-wave (DT) were measured and the E/A-ratio was calculated.

Tissue Doppler imaging: Color TDI tracings were obtained with the range gate placed at the septal

and lateral mitral annular segments in the 4-chamber view. The peak longitudinal early diastolic (e’)

velocity was measured and the average was calculated from the lateral and septal velocities and

used to obtain the E/e’.

The cardiac time intervals were obtained by placing a 2-4 cm straight M-mode line through the

7

septal half of the mitral leaflet in the color TDI 4-chamber view, and the time intervals were

measured directly from the color diagram1,6

(Figure 1). The IVCT was defined as the time interval

from the mitral valve (MV) closure (MVC), determined by the color shift from blue/turquoise to red

at end-diastole, to the aortic valve opening (AVO) determined by the color shift from blue to red

(Figure 1). The ET was defined as the time interval from the AVO to the aortic valve closing

(AVC), determined by the color shift from red to blue at end systole (Figure 1). The IVRT was

defined as the time interval from the AVC to the MV opening (MVO), determined by the color shift

from red-orange to yellow (Figure 1). The method has previously been validated 1,6,14

, and we have

previously demonstrated high reproducibility of the measure in this population1. Both isovolumic

time intervals were divided with ET creating IVRT/ET and IVCT/ET, respectively, and MPI was

calculated as the sum of the two ((IVRT+IVCT)/ET).

Follow-up and end Outcome: The primary endpoint of major cardiovascular events (MACE) was

the combined endpoint of being admitted with ischemic heart disease (IHD), admitted due to heart

failure (CHF) or cardiovascular mortality. Follow-up was 100%. Follow-up data on admission with

IHD and CHF were obtained from the Danish National Board of Health’s National Patient Registry,

using ICD-10 codes (DI20-DI259 for IHD and DI500-DI509 and DJ818 for CHF). Follow-up data

on cardiovascular mortality was collected from the national Danish Causes of Death Registry.

Cardiovascular death was defined as ICD-10 codes I00-I99.

Statistics: Proportions were compared using χ2-test, continuous Gaussian distributed variables with

Student’s t-test and Mann-Whitney test if non-Gaussian distributed.

Cumulative incidence curves of future MACE were obtained by competing risk Cox proportional

hazards regression models. Subdistribution Hazard Ratios (SHR) were calculated by competing risk

8

Cox proportional hazards regression analyses. Non-Gaussian distributed continues variables (LVMI

and E/e’) were categorised into quartiles when included in the models. The assumptions of

proportional hazards in the models were tested graphically and tested based on the Schoenfeld

residuals. Harrell’s c-statistics20

obtained from univariable Cox proportional hazards regression

models were calculated in order to test the prognostic performance of all the cardiac time intervals.

Furthermore, the Harrell’s c-statistics were obtained from multivariable Cox proportional hazards

regression models including all the conventional echocardiographic parameters and compared with

Harrell’s c-statistics obtained from multivariable Cox proportional hazards regression models

including all the conventional echocardiographic parameters and the IVRT/ET or MPI, respectively.

Predictive models, using logistic regression models, predicting the risk of future MACE, for

participants followed for median 10.8 years, were constructed. Reclassification analysis using

arbitrary risk categories of 0% to 5%, 5% to 10%, 10% to 20% and >20% was used to assessed the

integrated diagnostic improvement (IDI) and net reclassification index (NRI) when adding the

IVRT/ET, the MPI, LVEF<50%, E/e’, LAD, E/A-ratio or the LVMI to the significant clinical

predictors of MACE in our population. A p-value ≤ 0.05 in 2-sided test was considered statistically

significant. All analyses were performed with STATA Statistics/Data analysis, SE 12.0 (StataCorp,

Texas,USA).

9

Results

During follow-up (median 10.8 years, IQR: 8.6-11.5 years), 267 (13.9%) were admitted with IHD,

137 (7.2%) were admitted to hospital due to CHF and 134 (7.0%) participants died due to

cardiovascular causes. The combined endpoint was reached by 383 (20.0%) participants.

Baseline characteristics are displayed in Table 1 and 2.

The cardiac time intervals and prognosis: The risk of future IHD, being admitted with CHF or

dying due to cardiovascular causes, increased with increasing tertile of the IVRT (Figure 2a) and

the IVCT (Figure 2b), being approximately two times as high in the third tertile compared to the

first tertile (IVRT: SHR 2.4; 95% CI 1.9 to 3.1; p<0.001; IVCT: SHR 1.6; 95% CI 1.2 to 2.0;

p<0.001). For the ET the risk of future MACE increased with decreasing tertile of the ET (Figure

2c), being approximately two times as high in the first tertile compared to the third tertile (ET: SHR

1.5; 95% CI 1.2 to 2.0; p<0.001). After multivariable adjustment for all other clinical predictors of

MACE in our population, declining cardiac function, determined by prolongation of the IVRT and

the IVCT, remained independent predictors of MACE (Table 3). However, when adding the

conventional echocardiographic parameters to the model, none of the cardiac time intervals

remained independent predictors of MACE (Table 3).

The combined indexes obtained by combining the cardiac time interval and prognosis:

Impaired cardiac function determined by all the indexes obtained by combining the cardiac time

intervals IVRT/ET, IVCT/ET and MPI were associated with an increased risk of MACE (Figure 3).

The risk of MACE increased with increasing tertile of the combined indexes (Figure 3), being

approximately two times as high for the IVCT/ET and approximately three times as high for the

IVRT/ET and MPI in the third tertile compared to the first tertile (IVRT/ET: SHR 2.7, 95% CI 2.1

10

to 3.5, p<0.001; IVCT/ET: SHR 1.8, 95% CI 1.4 to 2.3, p<0.001; MPI: SHR 2.8, 95% CI 2.2 to 3.7,

p<0.001). After multivariable adjustment for both clinical predictors and conventional

echocardiographic measures, only the combined indexes, including information on both the systolic

and diastolic performance (the IVRT/ET and the MPI) remained independent predictors of future

MACE (Table 3).

Added value of the IVRT/ET and the MPI in relation to predicting future MACE:

Adding the combined indexes, including information on both the systolic and diastolic performance,

the IVRT/ET or the MPI, to a Cox model already including all other echocardiographic parameters

(LVEF<50%, E/e’, E/A-ratio, Dec time, LA dimension, LVMI) resulted in a significant increase in

the Harrell’s c-statistics (Figure 4a and b). Furthermore, reclassification analysis when adding the

combined indexes (the IVRT/ET or the MPI), to our clinical predictors (age, gender, BMI, eGFR,

Heart Rate, hypertension, diabetes, smoking status and atrial fibrillation), yielded better predicting

models with a significant increase in the categorical net reclassification index (NRI) of 3.052%

(95% CI 0.957-5.145%, p=0.004) for the IVRT/ET and 3.058% (95% CI 0.159-5.957%, p=0.039)

for the MPI, respectively (Table 4 and 5). In addition, when using continuous NRI, without using

arbitrary risk categories, the predicting models were improved even further, corresponding with an

increase in the continuous NRI of 19.2% (95% CI 3.2-33.0%) for the IVRT/ET and of 16.0% (95%

CI 5.5-33.5%) for the MPI, respectively. In comparison, none of the conventional

echocardiographic parameters evaluating systolic or diastolic function (LVEF<50%, E/e’, LAD or

E/A-ratio) or the LVMI yielded better predicting models when added to our clinical predictors

(LVEF<50%: NRI 1.384% (95% CI -0.035-2.803%), p=0.06; E/e’: NRI 2.803% (95% CI -0.185-

5.389%), p=0.07; LAD: NRI 1.883% (95% CI -0.554-4.300%), p=0.13; E/A-ratio: NRI -0.088%

(95% CI -2.014-1.839%), p=0.93; LVMI: NRI 0.968% (95% CI -3.409-5.345%), p=0.66).

11

Discussion

In the largest prospective study to date, of a random sample of participants from the general

population undergoing comprehensive echocardiography including an assessment of the cardiac

time intervals by TDI M-mode, we demonstrate: - Impaired cardiac time intervals are predictors of

future IHD, CHF or cardiovascular mortality in the general population. - The combined indexes

containing information on both systolic and diastolic function (the IVRT/ET and the MPI), are

independent predictors of future MACE, even after adjustment for all other echocardiographic

predictors. – The IVRT/ET and the MPI provides prognostic information in the general population,

not only independently, but also over and above all significant clinical and conventional

echocardiographic predictors of future MACE.

The novel method of obtaining the global cardiac time intervals: Cardiac intervals have been

suggested and investigated as potential markers of cardiac dysfunction for several decades21–24

.

However, the main concern has always been how to obtain these cardiac time intervals in a fast,

easy, non-invasive and reproducible manner. Tei and colleagues proposed in 1995 to overcome this

problem by obtaining the cardiac time intervals from the velocity curves obtained from pulsed-

Doppler echocardiography of the LV MV inflow and outflow tract and hereby calculating the index

of combined systolic and diastolic performance, the MPI23,24

. The method suggested by Tei and

colleagues has several limitations. Firstly, the time intervals are obtained from two projections and

at least two cardiac cycles23,24

, and may therefore be prone to changes in the cardiac time intervals

caused by heart rate variations in-between obtaining the two projections. Secondly, the IVRT and

IVCT are not obtained when using this method, only the sum of the IVRT and the IVCT23,24

. Using

TDI and the velocity curves from the myocardium is an alternative non-invasive approach to obtain

the cardiac time intervals2,3,25–27

. With this method the time intervals can be obtained from one

12

projection and one cardiac cycle. However, the time intervals obtained from the TDI velocity curves

are prone to regional differences in myocardial activation, mechanics and physiology28

. These

regional differences in the cardiac time intervals will be attenuated in persons suffering from, e.g.

bundle branch blocks, ischemic heart disease or LV dyssynchrony. However, we can overcome this

problem by analysing the global time intervals through evaluating the MV movement by a simple

color TDI M-mode analysis29

, instead of the regional velocity curves. Thus, using color TDI M-

mode through the mitral leaflet to estimate the cardiac time intervals is an improved method

reflecting global cardiac time intervals and eliminating beat-to-beat variation and regional

differences. We have previously demonstrated that the MPI obtained by TDI M-mode is less

influenced by physical parameters (including heart rate) compared to the conventional method

described by Tei and colleagues1. Hence, MPI based on simple cardiac mechanisms, i.e. the

passively opening and closing of the MV, is less confounded than MPI based on calculations from

versatile blood flow patterns through the mitral and aorta valve. Furthermore, the precision and

reproducibility of the cardiac time intervals and MPI are improved when they are obtained by the

TDI M-mode method compared to the conventional method1,14

. Additionally, it is not possible to

obtain the cardiac time intervals from velocity curves in patients with atrial fibrillation regardless if

they are obtained by the conventional method as describe by Tei and colleagues23,24

or by the newer

regional TDI velocity curve method25,26

. This is due to the absence of the A and a’ velocity curves

in patients with atrial fibrillation, which are needed to obtain the time intervals by both methods. In

contrast, all cardiac time intervals can be obtained by the novel color TDI M-mode method

regardless of the presence of atrial fibrillation6.

Also, the time intervals achieved by clear color shifts marking minimal changes in direction in the

MV by the color TDI M-mode method are very easy to identify (Figure 1). In contrast the time

intervals obtained by the conventional method23,24

or the newer regional TDI velocity curve

13

method25,26

are assessed from velocity curves where the signals may be scattered making it hard to

accurately define the cardiac time intervals, which is illustrated by the poorer reproducibility1,14

.

Furthermore, when good imaging quality is difficult to obtain, it is often possible to visualize the

MV in the apical view, due to the perpendicular nature between the ultrasound beam and the MV,

why assessing its longitudinal movement by color TDI M-mode nearly always is manageable.

The global cardiac time intervals and risk of future MACE: Previous studies, which evaluated

the prognostic value of the MPI either obtained by the conventional method as proposed by Tei and

colleagues23,24

or by using the regional TDI velocity curves25,26

, were all performed in selected

populations, e.g., patients after acute myocardial infarction30,31

, elderly men32

, patients with cardiac

amyloidosis33

, with idiopathic-dilated cardiomyopathy34

, with isolated diastolic dysfunction11

, and

in patients with systolic heart failure35

. In contrast, our study population was randomly selected

from a low risk general population, where subtle cardiac impairment often will go unrecognized if

only using conventional echocardiographic parameters. In a population-based study of American

Indians36

, the authors found that the MPI obtained by the conventional method23,24

did not predict

future cardiovascular events. In accordance with this, we have previously demonstrated that the

MPI obtained by the conventional method23,24

did not predict all-cause mortality in the general

population when adjusted for significant risk factors. However, the MPI obtained by the novel color

TDI M-mode method, was an independent predictor of mortality after multivariable adjustment1.

This discrepancy in predictive capacity between the two methods of obtaining the cardiac time

intervals is probably due to the difference in the physical mechanisms of the two methods. When

using the conventional method the time intervals are derived from blood flow velocity curves,

whereas when using the M-mode method the time intervals are derived from analysing the MV

moving passively depending on shifts in pressure and blood flow between the left atrium and

14

ventricle. Therefore, the method by Tei and colleagues23,24

seems to be influenced by more physical

parameters than the M-mode method1. This makes the MPI obtained by TDI M-mode less complex

to interpret in a clinical setting. These differences may also explain why the MPI obtained by the

TDI M-mode method is more precise and has better reproducibility1,14

. In accordance with this, MPI

obtained by using the regional TDI velocity curves25,26

has also demonstrated to be superior to the

conventional method23,24

in predicting cardiac mortality and admission with heart failure in patients

with isolated diastolic dysfunction or heart failure with preserved LVEF11

.

Our study is the first to evaluate the prognostic value of all the cardiac time intervals in a low risk

randomly selected general population. We found that impairment in all the cardiac time intervals

was individually associated with increased risk of MACE (Figure 2, 3 and Table 3). However, after

multivariable adjustment for all other clinical and echocardiographic predictors, only the combined

indexes containing information on both the systolic and diastolic performance (the IVRT/ET and

the MPI) remained independent predictors of future MACE (Table 3). We have previously observed

the same pattern in patients suffering a ST-segment elevation myocardial infarction (STEMI)

treated with primary percutaneous coronary intervention6. In this high risk population of STEMI

patients only the combined indexes containing information on both the systolic and diastolic

performance remained independent predictors of outcome after adjustment for all other clinical and

echocardiographic predictors6. The superiority of the IVRT/ET and the MPI compared to the

remaining cardiac time intervals in predicting MACE, may reflect that they detect myocardial

dysfunction, irrespective of myocardial sufferings from an ailing systolic or diastolic function6. In

contrast, the remaining cardiac time intervals only detect isolated diastolic (IVRT) or systolic

(IVCT, ET and IVCT/ET) dysfunction, respectively. Similar, when using TDI velocities in

predicting outcome, studies have demonstrated that when combining the information on systolic (s’)

15

and diastolic function (e’ and a’) in one risk stratification strategy, the prognostic information

obtained improves15,37,38

.

Incremental prognostic value of adding the IVRT/ET and the MPI in relation to predicting

future MACE: We found that the combined indexes containing information on both the systolic

and diastolic performance (the IVRT/ET and the MPI) remained not only independent predictors of

future MACE, but provided incremental prognostic value to all other echocardiographic predictors

of MACE (Figure 4). In addition, reclassification analysis, using arbitrary risk categories, when

adding the combined indexes (the IVRT/ET or the MPI) to our clinical predictors, yielded better

predicting models with significant increases in the categorical NRI of approximately 3.1% for both

the IVRT/ET and the MPI. In addition, when using continuous NRI, without using arbitrary risk

categories, the predicting models were improved even further, corresponding with an increase in the

continuous NRI of 19.2% for the IVRT/ET and of 16.0% for the MPI, respectively. In comparison,

neither of the conventional echocardiographic parameters evaluating systolic or diastolic function

(LVEF<50%, E/e’, LAD or E/A-ratio) or the LVMI yielded better predicting models when added to

our clinical predictors. The superiority of the IVRT/ET and the MPI to all the conventional

echocardiographic parameters evaluating systolic or diastolic function, may again reflect that they

detect very miniscule impairment in the myocardial function, irrespective if the myocardium suffers

from an ailing systolic or diastolic function6. Therefore, the IVRT/ET and MPI identifies subtle

impairments in the cardiac function (both systolic and diastolic), that may be unnoticed by

conventional echocardiography12,13

, and can improve risk stratification strategies evaluating future

risk of fulminant cardiovascular disease over and above the traditional predictors, in the low risk

general population. In contrast, it seems that adding an assessment of the conventional

16

echocardiographic parameters to the clinical risk factors, does not improve the risk stratification

strategy to identify high risk people in the general population.

Limitations

The inhabitants of Denmark and the study population are primarily Caucasian, which limits the

generalizability of our findings to other general populations with other composition.

Conclusion

In the general population, the combined cardiac time intervals which include information on both

the systolic and diastolic function in one index (IVRT/ET and MPI) are powerful and independent

predictors of future major cardiovascular events. In addition, they provide prognostic information

over and above clinical variables and conventional echocardiographic measures of systolic and

diastolic function.

Funding Sources

This study was financially supported by the Faculty of Health Sciences, University of Copenhagen,

Denmark. The sponsor had no role in the study design, data collection, data analysis, data

interpretation, or writing of the manuscript.

Conflict of interest

None declared

17

References

1. Biering-Sørensen T, Mogelvang R, Pedersen S, Schnohr P, Sogaard P, Jensen JS. Usefulness

of the myocardial performance index determined by tissue Doppler imaging m-mode for

predicting mortality in the general population. Am J Cardiol. 2011;107:478–483.

2. Gaibazzi N, Petrucci N, Ziacchi V. Left ventricle myocardial performance index derived either

by conventional method or mitral annulus tissue-Doppler: a comparison study in healthy

subjects and subjects with heart failure. J Am Soc Echocardiogr Off Publ Am Soc

Echocardiogr. 2005;18:1270–1276.

3. Duzenli MA, Ozdemir K, Aygul N, Soylu A, Aygul MU, Gök H. Comparison of myocardial

performance index obtained either by conventional echocardiography or tissue Doppler

echocardiography in healthy subjects and patients with heart failure. Heart Vessels.

2009;24:8–15.

4. Bruch C, Schmermund A, Marin D, Katz M, Bartel T, Schaar J, Erbel R. Tei-index in patients

with mild-to-moderate congestive heart failure. Eur Heart J. 2000;21:1888–1895.

5. Carluccio E, Biagioli P, Alunni G, Murrone A, Zuchi C, Biscottini E, Lauciello R, Pantano P,

Gentile F, Nishimura RA, Ambrosio G. Improvement of myocardial performance (Tei) index

closely reflects intrinsic improvement of cardiac function: assessment in revascularized

hibernating myocardium. Echocardiogr Mt Kisco N. 2012;29:298–306.

6. Biering-Sørensen T, Mogelvang R, Søgaard P, Pedersen SH, Galatius S, Jørgensen PG, Jensen

JS. Prognostic Value of Cardiac Time Intervals by Tissue Doppler Imaging M-Mode in

Patients With Acute ST-Segment-Elevation Myocardial Infarction Treated With Primary

Percutaneous Coronary Intervention. Circ Cardiovasc Imaging. 2013;6:457–465.

18

7. Su H-M, Lin T-H, Voon W-C, Lee K-T, Chu C-S, Yen H-W, Lai W-T, Sheu S-H. Correlation

of Tei index obtained from tissue Doppler echocardiography with invasive measurements of

left ventricular performance. Echocardiogr Mt Kisco N. 2007;24:252–257.

8. Uemura K, Kawada T, Zheng C, Li M, Shishido T, Sugimachi M. Myocardial performance

index is sensitive to changes in cardiac contractility, but is also affected by vascular load

condition. Conf Proc Annu Int Conf IEEE Eng Med Biol Soc IEEE Eng Med Biol Soc Conf.

2013;2013:695–698.

9. Su H-M, Lin T-H, Voon W-C, Lee K-T, Chu C-S, Lai W-T, Sheu S-H. Differentiation of left

ventricular diastolic dysfunction, identification of pseudonormal/restrictive mitral inflow

pattern and determination of left ventricular filling pressure by Tei index obtained from tissue

Doppler echocardiography. Echocardiogr Mt Kisco N. 2006;23:287–294.

10. Kraigher-Krainer E, Shah AM, Gupta DK, Santos A, Claggett B, Pieske B, Zile MR, Voors

AA, Lefkowitz MP, Packer M, McMurray JJV, Solomon SD, PARAMOUNT Investigators.

Impaired systolic function by strain imaging in heart failure with preserved ejection fraction. J

Am Coll Cardiol. 2014;63:447–456.

11. Kim H, Yoon H-J, Park H-S, Cho Y-K, Nam C-W, Hur S-H, Kim Y-N, Kim K-B. Usefulness

of tissue Doppler imaging-myocardial performance index in the evaluation of diastolic

dysfunction and heart failure with preserved ejection fraction. Clin Cardiol. 2011;34:494–499.

12. Ruggiero A, De Rosa G, Rizzo D, Leo A, Maurizi P, De Nisco A, Vendittelli F, Zuppi C,

Mordente A, Riccardi R. Myocardial performance index and biochemical markers for early

detection of doxorubicin-induced cardiotoxicity in children with acute lymphoblastic

leukaemia. Int J Clin Oncol. 2013;18:927–933.

19

13. Varol E, Akcay S, Ozaydin M, Ozturk O, Cerci SS, Sahin U. Influence of obstructive sleep

apnea on left ventricular mass and global function: sleep apnea and myocardial performance

index. Heart Vessels. 2010;25:400–404.

14. Kjaergaard J, Hassager C, Oh JK, Kristensen JH, Berning J, Sogaard P. Measurement of

cardiac time intervals by Doppler tissue M-mode imaging of the anterior mitral leaflet. J Am

Soc Echocardiogr Off Publ Am Soc Echocardiogr. 2005;18:1058–1065.

15. Mogelvang R, Sogaard P, Pedersen SA, Olsen NT, Marott JL, Schnohr P, Goetze JP, Jensen

JS. Cardiac dysfunction assessed by echocardiographic tissue Doppler imaging is an

independent predictor of mortality in the general population. Circulation. 2009;119:2679–

2685.

16. Langsted A, Freiberg JJ, Nordestgaard BG. Fasting and nonfasting lipid levels: influence of

normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction.

Circulation. 2008;118:2047–2056.

17. Rowley KG, Daniel M, O’Dea K. Screening for diabetes in Indigenous populations using

glycated haemoglobin: sensitivity, specificity, post-test likelihood and risk of disease. Diabet

Med J Br Diabet Assoc. 2005;22:833–839.

18. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW,

Materson BJ, Oparil S, Wright JT Jr, Roccella EJ, Joint National Committee on Prevention,

Detection, Evaluation, and Treatment of High Blood Pressure. National Heart, Lung, and

Blood Institute, National High Blood Pressure Education Program Coordinating Committee.

Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and

Treatment of High Blood Pressure. Hypertension. 2003;42:1206–1252.

20

19. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH,

Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MSJ, Stewart WJ.

Recommendations for chamber quantification: a report from the American Society of

Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification

Writing Group, developed in conjunction with the European Association of Echocardiography,

a branch of the European Society of Cardiology. J Am Soc Echocardiogr Off Publ Am Soc

Echocardiogr. 2005;18:1440–1463.

20. Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing

models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med.

1996;15:361–387.

21. Mancini GB, Costello D, Bhargava V, Lew W, LeWinter M, Karliner JS. The isovolumic

index: a new noninvasive approach to the assessment of left ventricular function in man. Am J

Cardiol. 1982;50:1401–1408.

22. Weissler AM, Harris WS, Schoenfeld CD. Systolic time intervals in heart failure in man.

Circulation. 1968;37:149–159.

23. Tei C. New non-invasive index for combined systolic and diastolic ventricular function. J

Cardiol. 1995;26:135–136.

24. Tei C, Ling LH, Hodge DO, Bailey KR, Oh JK, Rodeheffer RJ, Tajik AJ, Seward JB. New

index of combined systolic and diastolic myocardial performance: a simple and reproducible

measure of cardiac function--a study in normals and dilated cardiomyopathy. J Cardiol.

1995;26:357–366.

21

25. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Novel approach to measure myocardial

performance index: pulsed-wave tissue Doppler echocardiography. Echocardiogr Mt Kisco N.

2003;20:503–510.

26. Tekten T, Onbasili AO, Ceyhan C, Unal S, Discigil B. Value of measuring myocardial

performance index by tissue Doppler echocardiography in normal and diseased heart. Jpn

Heart J. 2003;44:403–416.

27. Harada K, Tamura M, Toyono M, Oyama K, Takada G. Assessment of global left ventricular

function by tissue Doppler imaging. Am J Cardiol. 2001;88:927–932, A9.

28. Cui W, Roberson DA, Chen Z, Madronero LF, Cuneo BF. Systolic and diastolic time intervals

measured from Doppler tissue imaging: normal values and Z-score tables, and effects of age,

heart rate, and body surface area. J Am Soc Echocardiogr Off Publ Am Soc Echocardiogr.

2008;21:361–370.

29. Voigt J-U, Lindenmeier G, Exner B, Regenfus M, Werner D, Reulbach U, Nixdorff U,

Flachskampf FA, Daniel WG. Incidence and characteristics of segmental postsystolic

longitudinal shortening in normal, acutely ischemic, and scarred myocardium. J Am Soc

Echocardiogr Off Publ Am Soc Echocardiogr. 2003;16:415–423.

30. Møller JE, Egstrup K, Køber L, Poulsen SH, Nyvad O, Torp-Pedersen C. Prognostic

importance of systolic and diastolic function after acute myocardial infarction. Am Heart J.

2003;145:147–153.

31. Møller JE, Søndergaard E, Poulsen SH, Egstrup K. The Doppler echocardiographic

myocardial performance index predicts left-ventricular dilation and cardiac death after

myocardial infarction. Cardiology. 2001;95:105–111.

22

32. Arnlöv J, Ingelsson E, Risérus U, Andrén B, Lind L. Myocardial performance index, a

Doppler-derived index of global left ventricular function, predicts congestive heart failure in

elderly men. Eur Heart J. 2004;25:2220–2225.

33. Tei C, Dujardin KS, Hodge DO, Kyle RA, Tajik AJ, Seward JB. Doppler index combining

systolic and diastolic myocardial performance: clinical value in cardiac amyloidosis. J Am

Coll Cardiol. 1996;28:658–664.

34. Dujardin KS, Tei C, Yeo TC, Hodge DO, Rossi A, Seward JB. Prognostic value of a Doppler

index combining systolic and diastolic performance in idiopathic-dilated cardiomyopathy. Am

J Cardiol. 1998;82:1071–1076.

35. Correale M, Totaro A, Greco CA, Musaico F, De Rosa F, Ferraretti A, Ieva R, Di Biase M,

Brunetti ND. Tissue Doppler time intervals predict the occurrence of rehospitalization in

chronic heart failure: data from the daunia heart failure registry. Echocardiogr Mt Kisco N.

2012;29:906–913.

36. Mishra RK, Kizer JR, Palmieri V, Roman MJ, Galloway JM, Fabsitz RR, Lee ET, Best LG,

Devereux RB. Utility of the myocardial performance index in a population with high

prevalences of obesity, diabetes, and hypertension: the strong heart study. Echocardiogr Mt

Kisco N. 2007;24:340–347.

37. Biering-Sørensen T, Jensen JS, Pedersen S, Galatius S, Hoffmann S, Jensen MT, Mogelvang

R. Doppler Tissue Imaging Is an Independent Predictor of Outcome in Patients with ST-

Segment Elevation Myocardial Infarction Treated with Primary Percutaneous Coronary

Intervention. J Am Soc Echocardiogr Off Publ Am Soc Echocardiogr. 2014;27:258–267.

23

38. Andersson C, Gislason GH, Møgelvang R, Hoffmann S, Mérie C, Køber L, Torp-Pedersen C,

Søgaard P. Importance and inter-relationship of tissue Doppler variables for predicting adverse

outcomes in high-risk patients: an analysis of 388 diabetic patients referred for coronary

angiography. Eur Heart J Cardiovasc Imaging. 2012;13:643–649.

24

Figure legends

Figure 1: Title: The cardiac time intervals assessed by a color TDI M-mode line through the mitral

leaflet.

Caption: Left: Four-chamber gray-scale (bottom) and color TDI (top) views in end-systole

displaying the position of the M-mode line used for measuring the cardiac time intervals. Right:

Color diagram of the TDI M-mode line through the mitral leaflet.

MVC=MV Closing; AVO=Aortic Valve Opening; AVC=Aortic Valve Closure; MVO=MV

Opening.

Figure 2: Title: Cardiac time intervals and outcome.

Caption: Figure 2a: Depicting the cumulative incidence of MACE for participants stratified

according to the tertiles of the IVCT (1.tertile < 30 ms; 2.tertile ≥ 30 ms to < 40 ms; 3.tertile ≥ 40

ms). Figure 2b: Depicting the cumulative incidence of MACE for participants stratified according

to the tertiles of the IVRT (1.tertile < 90 ms; 2.tertile ≥ 90 ms to < 109 ms; 3.tertile ≥ 109 ms).

Figure 2c: Depicting the cumulative incidence of MACE for participants stratified according to the

tertiles of the ET (1.tertile < 275 ms; 2.tertile ≥ 275 ms to < 296 ms; 3.tertile ≥ 296 ms). IVCT =

Isovolumic Contraction Time; IVRT = Isovolumic Relaxation Time; ET = Ejection Time

Figure 3: Title: The combined indexes of the cardiac time intervals and outcome.

Caption: Figure 3a: Depicting the cumulative incidence of MACE for participants stratified

according to the tertiles of the combined index obtained by dividing the IVCT by the ET (1.tertile <

0.10; 2.tertile ≥ 0.10 to < 0.14; 3.tertile ≥ 0.14). Figure 3b: Depicting the cumulative incidence of

MACE for participants stratified according to the tertiles of the combined index obtained by

dividing the IVRT by the ET (1.tertile < 0.31; 2.tertile ≥ 0.31 to < 0.39; 3.tertile ≥ 0.39). Figure 3c:

25

Depicting the cumulative incidence of MACE for participants stratified according to the tertiles of

the MPI (1.tertile < 0.43; 2.tertile ≥ 0.43 to < 0.52; 3.tertile ≥ 0.52). IVCT = Isovolumic Contraction

Time; IVRT = Isovolumic Relaxation Time; ET = Ejection Time; MPI = Myocardial Performance

Index.

Figure 4: Title: Incremental prognostic information by adding an assessment of the combined

cardiac time intervals (the IVRT/ET and the MPI) to the conventional echocardiographic measures.

Caption: The Harrell's c-statistic values obtained from the multivariable Cox proportional hazards

regression models. Conventional measures of systolic and diastolic function includes: LVEF<50%,

E/e’, E/A-ratio, Dec time, LA dimension. Conventional measures of systolic and diastolic function,

and the left ventricular mass index includes: LVEF<50%, E/e’, E/A-ratio, Dec time, LA dimension

and LVMI.

26

Table 1. Baseline clinical characteristics for the participants stratified according to major adverse

cardiovascular (MACE) outcome

BMI=Body Mass Index; eGFR= Estimated Glomerular Filtration Rate

No MACE

(n=1532)

MACE

(n=383) P-value

Age (years)

55±16 71±11 <0.001

Male gender

41% 49% 0.006

Systolic Blood Pressure (mmHg) 133±22 148±21 <0.001

Diastolic Blood Pressure (mmHg) 78±12 81±13 <0.001

Heart Rate (beats per minute) 66±11 70±12 <0.001

Hypertension 36% 73% <0.001

Diabetes

8% 19% <0.001

Smoking status

- Never

- Previous

- Current

36%

31%

33%

25%

42%

34%

<0.001

Cholesterol 5.5±1.2 5.6±1.1 0.25

Previous ischemic heart disease

2% 23% <0.001

Previous ischemic stroke 1% 5% <0.001

BMI (kg/m2) 25.1±3.8 26.9±4.1 <0.001

eGFR (mL/min/1,73 m2) 77.2±15.5 72.4±18.0 <0.001

Atrial fibrillation/flutter 1% 6% <0.001

27

Table 2. Baseline echocardiographic characteristics for the participants stratified according to major

adverse cardiovascular outcome (MACE)

LVEF=Left Ventricular Ejection Fraction; LVMI=Left Ventricular Mass Index; E=peak transmitral

early diastolic inflow velocity; A=peak transmitral late diastolic inflow velocity; DT=deceleration

time of early diastolic inflow; e’=average peak early diastolic longitudinal mitral annular velocity

determined by color TDI; IVCT=Isovolumic Contraction Time; IVRT=Isovolumic Relaxation

Time; ET=Ejection Time; MPI = Myocardial Performance Index.

No MACE

(n=1463)

MACE

(n=383) P-value

LVEF < 50% 0% 4% <0.001

LVIDd (cm) 4.8±0.5 4.8±0.7 0.91

LAD (cm) 3.4±0.4 3.6±0.5 <0.001

E (m/s) 0.73±0.16 0.69±0.18 <0.001

A (m/s) 0.67±0.18 0.76±0.21 <0.001

E/A ratio < 1 37% 61% <0.001

DT (ms) 165±37 174±49 <0.001

E/e’

9.5 (7.9-11.8) 12.6 (10.2-16.6) <0.001

LVMI

80.9 (70.1-95.1) 94.7 (79.2-113.2) <0.001

Cardiac time intervals:

IVRT (ms) 98±22 111±29 <0.001

IVCT (ms) 36±13 40±16 <0.001

ET (ms) 286±25 277±31 <0.001

IVRT/ET 0.35±0.08 0.41±0.14 <0.001

IVCT/ET

0.13±0.05 0.15±0.07 <0.001

MPI 0.47±0.11 0.56±0.18 <0.001

28

Table 3. Unadjusted and adjusted a competing risk Cox proportional hazards regression models depicting the cardiac time intervals as

predictors of future major adverse cardiovascular outcome (MACE)

Harrell’s

c-statistics

Univariable

SHR (95% CI)

P-value

Model 1

Multivariable

Adjustment

SHR (95% CI)

P-value

Model 2

Multivariable

Adjustment

SHR (95% CI)

P-value

IVRT per 10 ms

increase

0.62 1.20 (1.16-1.25) <0.001 1.06 (1.02-1.12) 0.004 1.07 (1.00-1.14) 0.051

IVCT per 10 ms

increase

0.57 1.19 (1.11-1.27) <0.001 1.10 (1.02-1.19) 0.010 1.07 (0.98-1.17) 0.14

ET per 10 ms

decrease

0.57 1.13 (1.08-1.18) <0.001 1.05 (0.99-1.11) 0.12 1.01 (0.94-1.07) 0.84

IVRT/ET per 0.1

increase

0.64 1.44 (1.31-1.58) <0.001 1.17 (1.08-1.27) <0.001 1.16 (1.00-1.34) 0.047

IVCT/ET per 0.1

increase

0.59 1.71 (1.46-2.02) <0.001 1.33 (1.10-1.60) 0.003 1.18 (0.96-1.45) 0.11

MPI per 0.1

increase

0.64 1.41 (1.33-1.50) <0.001 1.17 (1.09-1.25) <0.001 1.11 (1.03-1.23) 0.024

29

Model 1 is adjusted for age, gender, BMI, eGFR, Heart Rate, hypertension, diabetes, smoking status, atrial fibrillation, ischemic heart

disease, previous ischemic stroke. Model 2 is adjusted for the same variables as Model 1 and for systolic dysfunction determined by LVEF

< 50%, LA diameter, LVMI, DT, E/A-ratio < 1 and E/e’.

SHR=Subdistribution Hazard Ratio; BMI=Body Mass Index; eGFR= Estimated Glomerular Filtration Rate; LVEF = Left Ventricular

Ejection Fraction; LVMI = Left Ventricular Mass Index; E = peak transmitral early diastolic inflow velocity; A = peak transmitral late

diastolic inflow velocity; DT = deceleration time of early diastolic inflow; e’ = average peak early diastolic longitudinal mitral annular

velocity determined by color TDI; IVCT=Isovolumic Contraction Time; IVRT=Isovolumic Relaxation Time; ET=Ejection Time; MPI =

Myocardial Performance Index.

30

Table 4. Risk Classification Tables for adding the IVRT/ET to the prediction model

Baseline risk factors include age, gender, BMI, eGFR, Heart Rate, hypertension, diabetes, smoking

status and atrial fibrillation. BMI=Body Mass Index; eGFR= Estimated Glomerular Filtration Rate.

Classification tables of predicting risk from the baseline risk factors,

and from the baseline risk factors and IVRT/ET

Non-Cases Predicting risk from a model including the baseline

risk factors and IVRT/ET

Predicting risk from a

model including the

baseline risk factors

<5% 5%-10% 10-20% ≥20% Total

<5% 494 13 507

5-10% 20 208 15 243

10-20% 22 248 19 289

≥20% 23 407 430

Total 514 243 286 426 1,469

Cases Predicting risk from a model including the baseline

risk factors and IVRT/ET

Predicting risk from a

model including the

baseline risk factors

<5% 5%-10% 10-20% ≥20% Total

<5% 10 10

5-10% 14 2 16

10-20% 1 43 4 48

≥20% 3 298 301

Total 10 15 48 302 375

IDI 0.335% (95% CI -0.126-0.796%), p=0.15

Categorical NRI 3.052% (95% CI 0.957-5.145%), p=0.004

Continuous NRI 19.2% (95% CI 3.2-33.0%)

31

Table 5. Risk Classification Tables for adding the MPI to the prediction model

Baseline risk factors include age, gender, BMI, eGFR, Heart Rate, hypertension, diabetes, smoking

status and atrial fibrillation. BMI=Body Mass Index; eGFR= Estimated Glomerular Filtration Rate.

Classification tables of predicting risk from the baseline risk factors,

and from the baseline risk factors and MPI

Non-Cases Predicting risk from a model including the baseline

risk factors and MPI

Predicting risk from a

model including the

baseline risk factors

<5% 5%-10% 10-20% ≥20% Total

<5% 488 19 507

5-10% 20 203 20 243

10-20% 31 227 31 289

≥20% 29 401 430

Total 508 253 276 432 1,469

Cases Predicting risk from a model including the baseline

risk factors and MPI

Predicting risk from a

model including the

baseline risk factors

<5% 5%-10% 10-20% ≥20% Total

<5% 8 2 10

5-10% 14 2 16

10-20% 3 37 8 48

≥20% 9 293 301

Total 8 19 47 301 337

IDI 0.724% (95% CI 0.079-1.252%), p=0.028

Categorical NRI 3.058% (95% CI 0.159-5.957%), p=0.039

Continuous NRI 16.0% (95% CI 5.5-33.5%)

32

Figure 1

33

Figure 2

0

.05

.1.1

5.2

.25

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the IVCT

0.1

.2.3

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the IVRT

0

.05

.1.1

5.2

.25

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the ET

Figure 2a.

Figure 2b.

Figure 2c.

34

Figure 3.

0.1

.2.3

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the IVCT/ET

0.1

.2.3

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the IVRT/ET

0.1

.2.3

Cum

ula

tive In

cid

en

ce

0 5 10 15Years

1. tertile 2. tertile 3. tertile

MACE by tertiles of the MPI

Figure 3a.

Figure 3b.

Figure 3c.

35

Figure 4a.

Figure 4b.

0.73

0.74

0.75

0.76

0.77

0.78

Conventional measuresof systolic and diastolic

function

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index +the IVRT/ET

0.7530148

0.758623

0.7667751

P=0.015

P=0.15

P=0.003

0.73

0.74

0.75

0.76

0.77

0.78

Conventional measuresof systolic and diastolic

function

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index

Conventional measuresof systolic and diastolicfunction, and the left

ventricular mass index +MPI

0.7530148

0.758623

0.7644389

P=0.029

P=0.15

P=0.010

Harr

el’s

c-s

tatistics

Harr

el’s

c-s

tatistics