Markers of inflammation and haemostasis: Associations with myocardial injury, adverse remodelling and future clinical events in patients with ST- elevation myocardial infarction PhD thesis Vibeke N. Ritschel, MD Center for Clinical Heart Research and Cardiac Intensive Care Research Unit, Department of Cardiology Oslo University Hospital Ullevaal & Faculty of Medicine University of Oslo, Oslo Norway
52
Embed
Markers of inflammation and haemostasis ... - duo.uio.no
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Markers of inflammation and haemostasis: Associations with myocardial injury, adverse remodelling and future clinical events in patients with ST-
elevation myocardial infarction
PhD thesis
Vibeke N. Ritschel, MD
Center for Clinical Heart Research and Cardiac Intensive Care Research Unit, Department of Cardiology
The atherosclerotic process and plaque rupture.............................................................................12Haemostasis, coronary thrombosis and fibrinolysis......................................................................13
Inflammation.............................................................................................................................15Interleukin-6 signalling.................................................................................................................15Myocardial injury, remodelling and function................................................................................18CCN2/connective tissue growth factor (CTGF)............................................................................19
Materialandmethods......................................................................................................21Study subjects and design..............................................................................................................21Coronary angiography and PCI.....................................................................................................22Ischemic postconditioning (iPost).................................................................................................22Blood sampling and laboratory analyses.......................................................................................23Echocardiography..........................................................................................................................24Cardiac magnetic resonance (CMR).............................................................................................24Endpoint registration.....................................................................................................................25Statistics........................................................................................................................................25
Members of the IL-6 family in patients with acute myocardial infarction....................................33Members of the IL-6 family and their association with later clinical events.................................35Inflammation as therapeutic target in patients with acute myocardial infarction..........................37Hypercoagulability in patients with acute myocardial infarction..................................................38CCN2 in patients with acute myocardial infarction......................................................................40
This project has been carried out at Center for Clinical Heart Research and Cardiac Intensive
Care Research Unit, Department of Cardiology, Oslo University Hospital Ullevaal during the
period of 2010 - 2017. I am grateful for the financial support from Stein Erik Hagen
Foundation for Clinical Heart Research.
I want to express my sincere gratitude to everyone who has contributed to this thesis. First of
all I would like to thank my main supervisor Geir Øystein Andersen who offered me the
position as a research fellow and introduced me to the clinical cardiovascular research. He
brought up the idea and formulated the hypothesis. Being a full-time cardiologist and giving
necessary supervision, continuous and constructive feedback has been essential for this work.
Second, I want to thank my co-supervisor, Ingebjørg Seljeflot, for her knowledge, experience,
constructive feedback and advices. She showed great patience and offered invaluable support
throughout all these demanding years. I am grateful for being a part of the research group led
by her and Harald Arnesen. I want to extend my gratitude to my second co-supervisor, Jan
Eritsland, for his excellent feedback, scientific knowledge and constructive responses.
Third, Harald Arnesen, Thomas Weiss, Sigrun Halvorsen, Limalanathan Shanmuganathan and
Pavel Hoffmann, all been co-authors of the present work, are appreciated for their great job in
order for this thesis to take its definite form.
Special thanks to Christian Shetelig for constructive collaboration and shared authorship of
the fourth article.
6
I am grateful to study nurse Charlotte Holst Hansen for her support and collaboration with the
third paper of this thesis. I also want to thank her for help with collecting blood samples and
clinical data for the BAMI biobank and database.
Special thanks to Morten Fagerland for giving me insights in statistical methods and for
guiding me through the statistical work in the second paper. Many thanks to Sissel Åkra for
her tremendous work with the laboratory analyses in this thesis, and to the study nurses and
the staff at the Intensive Cardiac Care Unit and Center for Clinical Heart Research for their
excellent assistance for the BAMI biobank and database.
Further, I want to thank my colleagues, fellow PhD students and seniors at the Center for
Clinical Heart Research and Department of Cardiology at OUS Ullevaal for their support.
Last but not least, I want to thank my husband Michael, my children Leonard and Catarina,
my parents Britt and Steinar and my brother Erlend and his wife Carmen for their patience,
continuous support and love. They helped me balance the Ph-D project and my clinical work
at NIMI during an intensive house-building period. They all are keystones in my life.
7
Abbreviations
ACS Acute coronary syndrome AKT Protein kinase B (serine/threonine-specific
protein kinase) AMI Acute myocardial infarction BAMI Biobanking in Acute Myocardial Infarction BMI Body mass index CAD Coronary artery disease CAT Calibrated automated thrombogram CCN2/CTGF CCN family protein 2/connective tissue
growth factor CHD Coronary heart disease CMR Cardiac magnetic resonance CRP C-reactive protein CVD Coronary vascular disease ECG Electrocardiogram ELISA Enzyme-linked immunosorbent assay ETP Endogenous thrombin potential ERK-kinase Extracellular signal regulated kinase F1+F2 Prothrombin fragment 1+2 HbA1c Glycosylated haemoglobin HT Hypertension IL-6 Interleukin-6 IR Ischemia-reperfusion JAK Janus kinase LDL Low-density lipoprotein LT Lag time LV Left ventricular LVEDVi Indexed left ventricular end-diastolic volume LVEF Left ventricular ejection fraction LVESVi Indexed left ventricular end-systolic volume MAP kinase Mitogen-activated protein kinase MVO Microvascular obstruction NT-proBNP N-terminal pro B-type natriuretic peptide PCI Percutaneous coronary intervention POSTEMI Postconditioning in ST-elevation myocardial
infarction pTG Peak thrombin generation P13k Phosphoinositide 3-kinase Q Quartile sgp130 Soluble glycoprotein 130 sIL-6R Soluble interleukin-6 receptor STAT Signal transducer and activator of
transcription family STEMI ST- elevation myocardial infarction TF Tissue factor TGFß1 Transforming growth factor beta 1 TnT Troponin T
Eritsland J, Fagerland MW, Andersen GØ. Evaluation of circulating levels of
CCN2/connective tissue growth factor in patients with ST-elevation myocardial
infarction. Scientific Reports 2017;7:11945.
*Shared authorship
10
11
Introduction
Cardiovascular disease and acute myocardial infarction
The cardiovascular mortality rate is reduced by approximately 50 % in the Western world
during the last 30 years1, 2, but cardiovascular disease (CVD) is still the major cause of death
and remains a challenge to public health2. CVD will remain one of the most frequent causes
of death in the future3, and a huge increase is expected in Eastern countries. CVD is mainly
caused by an ongoing atherosclerotic and inflammatory process in the medium-sized and
large arteries 4, 5, and acute myocardial infarction (AMI) is together with sudden cardiac death
and stroke the most serious consequence of the atherosclerotic process6. Despite progress in
treatment, new clinical adverse events, including AMI, stroke, heart failure, repeated
revascularisation and death, are still frequent in survivors of AMI, thus, it is important to
identify novel risk factors associated with long-term prognosis.
AMI is usually initiated by endothelial erosion or plaque rupture which may lead to an
occluding coronary thrombus5. The consequence is ischemia leading to necrosis of the
affected myocardium7-9. AMI is characterised by symptoms of ischemia with rise and fall in
levels of specific cardiac markers, i.e. troponins with minimum one value above the 99th
percentile. This, together with new ST-segment changes, left bundle-branch block or Q wave
appearance in the ECG, loss of viable myocardium or regional hypokinesia/akinesia judged
by echocardiography or cardiac magnetic resonance (CMR) imaging, or signs of an acute
thrombus on the coronary angiogram or by autopsy, make up the definition of AMI10, 11. AMI
is divided into two groups according to ECG changes: ST-elevation myocardial infarction
(STEMI)11 and non-ST-elevation myocardial infarction (non-STEMI). STEMI may appear as
anterior, inferior, lateral, or posterior, depending on the location of the injured artery. Non-
STEMI is more common than STEMI, the frequency being about 70 % of all AMI12.
12
The atherosclerotic process and plaque rupture
The atherosclerotic process is associated with known common risk factors like hypertension,
smoking, elevated blood lipids and diabetes mellitus. Progression of the pathophysiological
changes in the vessel-wall13 may lead to abnormal blood flow, disturbed shear stress,
endothelial injury, plaque rupture and thrombus formation. Low-density lipoprotein (LDL) is
a crucial component in the development of atherosclerosis and plaque formation. In this
process LDL particles are modified into oxidized LDL (oxLDL) by free-radical oxygen
species. OxLDL particles accumulate in the vessel wall (Figure 1)4, 14 and induce a low-grade
inflammatory response in the arterial intima13, 15 and foam cell production. OxLDL particles
thus promotes the atherosclerotic plaque formation, which can lead to plaque rupture with
subsequent activation of platelets and the coagulation system with risk of thrombus
development and potential vessel occlusion 4, 13, 16. It is accepted that both the innate and
adaptive immune system are activated, involving neutrophils and monocyte recruitment,
cytokine activation, migration of smooth muscle cells and recruitment of T-lymphocytes
among others, and both systems may be involved in the plaque development16.
From a clinical point of view, the involvement of inflammatory mediators in the
atherosclerotic process raises the possibility of developing new therapeutic strategies.
13
Fig. 1 Simplified visualization of arterial plaque formation, rupture and thrombus formation. Adapted from Choudhury RP et al with permission. Copyright Clearance Center´s RightsLink ®service14
Haemostasis, coronary thrombosis and fibrinolysis
The haemostatic system balances platelet activation, coagulation and fibrinolysis in order to
maintain normal blood flow 13, 17, 18. Coagulation can be initiated through both the intrinsic
and extrinsic pathways19 (Figure 2), and is amplified by the phospholipid surface of activated
platelets. In the context of plaque rupture, platelets are activated and aggregated and in
addition to amplify thrombin generation and blood clotting also play a role in the
inflammatory response4. Tissue factor (TF), the most important activator of the extrinsic
coagulation cascade, when exposed to circulating blood, binds to and activate coagulation
factor FVII, with further activation of FX to Xa, initiating the conversion of pro-thrombin to
thrombin and further clot formation by converting fibrinogen to fibrin (Figure 2) 17, 19.
Increased thrombin generation has been shown in patients with AMI and is crucial in the
thrombus formation20-24.
14
By measuring split products from the conversion of prothrombin to thrombin (prothrombin
fragments F1+2), an ongoing in vivo thrombotic process can be estimated25. The endogenous
thrombin potential (ETP) can be used as an estimate of ex vivo potential to generate thrombin
(ref. page 23). Small amounts of fibrin formed, also initiate the fibrinolytic system, balancing
the haemostatic system. The conversion of plasminogen to plasmin by tissue plasminogen
activator (t-PA) results in cleavage of fibrin into fibrin degradation products, which can be
assessed by measurement of D-dimer25, reflecting an ongoing activation of both the
coagulation and fibrinolytic systems26. Like prothrombin fragments F1+2, D-dimer has been
shown to be elevated for a prolonged time in patients with AMI22 and an early reduction of D-
dimer is supposed to indicate better prognosis27
Figure 2. The coagulation and fibrinolytic systems. Reproduced from Kohler with permission. Copyright Massachusetts Medical Society ®17
15
Inflammation
The inflammatory process during and after an AMI is complex and depends on a variety of
cytokines and their receptors, acting together to initiate complex signalling cascades28, 29. The
inflammatory response is crucial for the repairing process in the injured myocardium30.
However, an excessive and prolonged inflammatory process can be harmful and may
contribute to larger infarct size, adverse remodelling and heart failure development.8 Studies
on circulating biomarkers like N-terminal pro-B-type natriuretic peptide (NT-pro-BNP) and
C-reactive protein (CRP) have identified important mechanistic insight about prognosis and
risk stratification31, 32. However, there are still many unknown mechanisms to explore. In the
present thesis members of the interleukin-6 (IL-6) transsignalling pathway, circulating IL-6,
growth factor (CCN2) and their importance in the early phase of an AMI as well as in long-
term prognosis, are focused.
Interleukin-6 signalling
The IL-6 family of cytokines signals through the common receptor unit gp130 and includes in
addition to IL-6, IL-11, leukaemia inhibitory factor, oncostatin M, ciliary neurotropic factor
and cardiotrophin-133.
IL-6, a multifunctional cytokine with both pro- and anti-inflammatory effects34,35, is released
from different cell types (monocytes/macrophages, B and T- lymphocytes, skeletal muscle
cells, endothelial cells, adipocytes, fibroblasts, vascular smooth muscle cells36, 37) and has a
variety of biological functions, especially in the acute-phase response38. IL-6 may be involved
both in the acute inflammatory response accompanying an AMI and in the chronic low-grade
systemic inflammation associated with cardiovascular events38-40.
16
IL-6 is produced in hypoxic cardiomyocytes29, 41, which may contribute to cardiac dysfunction
and ischemia-reperfusion (IR) injury28.
Elevated circulating levels of IL-6 after an AMI have also been reported 42, 43. It remains
unclear whether the IL-6 response related to myocardial necrosis is beneficial, detrimental or
both. IL-6 stimulates the production of C-reactive (CRP) from the liver36 and may be the link
between high levels of CRP and increased risk of cardiovascular events and worsening of
heart failure42, 44-50.
There are two different signalling pathways of IL-6, the classical and the transsignalling, and
three main intracellular signalling cascades are thought to be involved: the Jak/STAT1/3, the
ERK 1/2/MAP kinase and the P13K/AKT pathways36, 37, 51. In the classical signalling
pathway, IL-6 binds to the membrane-bound IL-6R, leading to activation of the early immune
response and possibly regenerative or anti-inflammatory processes 35, 36, 52, 53 (Figure 3A). The
classical membrane-bound IL-6R stimulated pathway is relatively specific due to the limited
number of cells (e.g. hepatocytes, white blood cells) that express IL-6Rs in the cell
membrane54. In the transsignalling pathway, in contrast to the IL-6R pathway, the cellular
ligand of the IL-6/sIL-6R complex on the cell surface, the receptor subunit glycoprotein 130
(gp130 protein- ß subunit) is widely expressed in most cells in the body. Shedding of the IL-6
receptor to the blood stream creates a soluble form of the IL-6 receptor (sIL-6R). sIL-6R
binds to circulating IL-6 and binds subsequently to gp130 on the cell surface and initiate
intracellular signalling mediated through the Jak/STAT1/3 cascade35, 55 (Figure 3 A). This
transsignalling pathway may be responsible for the pro-inflammatory effects of IL-635. The
soluble form of gp130 is the natural inhibitor of the sIL-6R/IL6 complex by binding to the
membrane bound gp130 (Figure 3B)35, 55, 56 and thereby reduce activation of the
transsignalling pathway.
17
Figure 3. The IL-6R complex, classical and transsignalling pathways, and the inhibitory mechanism of sgp130. (A) The two modes of IL-6 activation. (B) Classical signalling is unaffected by sgp130. Reproduced from Rose-John with permission. Copyright Clearance Center’s RightsLink® service55
The importance of IL-6 and CRP has been extensively studied and they are thought to play a
role in all stages of the atherosclerotic development from early lesions to destabilisation of the
plaque57. They are important for reperfusion injury, repair processes and scar tissue formation
after an AMI28 and both biomarkers have been shown to be associated with increased risk in
patients with CAD6, 37, 58 and with large infarct size in AMI patients59. In contrast, the soluble
receptors of the IL-6 family (sIL-6R and sgp130) have been sparsely studied in AMI patients.
18
Myocardial injury, remodelling and function
LV remodelling after an AMI alters ventricular size, shape and function and involve
mechanical, neurohormonal and possibly genetic factors.
Loss of cardiomyocytes and interstitial fibrosis following an AMI induce an inflammatory
response, which is important for the repair process of the injured myocardium, scar healing
and post-infarction remodelling38. Also, molecular, cellular and interstitial changes occur as a
consequence of myocardial cell death after an AMI. Abnormal loading conditions may be
followed by left or right ventricular dilatation, changes of the shape of the LV, as well as
hypertrophy60. The ventricles are prone to dilatation (eccentric remodelling), increased
hypertrophy and myocardial dysfunction after an AMI (Figure 4)38, 61, 62.
Figure 4. Remodelling after an acute AMI with scar tissue formation and dilatation of the left ventricle. Reproduced from Konstam with permission. Copyright Clearance Center´s Rights link service®62
Inflammatory cell activation is part of the first response after AMI63, 64 and together with
chemokines like IL-8 is thought to act chemotactic in the early phase of an AMI, by recruiting
neutrophils and monocytes to the infarct area30. This early response is followed by release of
19
different growth factors like CCN2/ connective tissue growth factor known to be involved in
cardiac hypertrophy and fibrosis63.
CCN2/connective tissue growth factor (CTGF)
CCN2/CTGF is a growth factor and member of the CCN family of matricellular proteins65.
Increased expression of CCN2 has been found in models of pressure-overload heart failure, 66,
67 and myocardial infarction68, 69 and is thought to be involved in collagen production leading
to myocardial fibrosis, scar formation and ventricular remodelling. Furthermore, CCN2 is
involved in extracellular matrix production and angiogenesis and exerts its effects on many
cell types involved in tissue repair, like neutrophils, fibroblasts, smooth muscle cells and
angiotensin II have all been shown to up-regulate CCN2 in different cell types after an AMI
71. In experimental studies CCN2 have mainly been seen as a pro-fibrotic marker. However,
results from experimental studies have also indicated that CCN2 may have cardioprotective
properties by attenuating LV hypertrophy72 and by improving tolerance to IR injury73. In a
small clinical study, high circulating levels of CCN2 were reported to be associated with
improved LV function74, however, there is a lack of data from clinical studies with adequate
sample size and long-term follow up, and the exact role of CCN2 in patients with AMI is still
unknown.
20
Aims of the thesis
The overall aims of the thesis were to improve our understanding of the complex mechanism
involved in the interaction between inflammation, thrombosis and remodelling in patients
with STEMI and, furthermore, to identify novel biomarkers and their possible association
with risk factors of CVD, myocardial injury, infarct size, adverse remodelling and long-term
clinical outcome.
Specific Aims
• To study the association between circulating levels of IL-6, sIL-6R, sgp130 and
CRP and 1) known risk factors of CVD and 2) myocardial injury and function in a
cohort of 1028 STEMI patients (Paper I)
• To investigate circulating levels of IL-6, sIL-6R, sgp130 and CRP as related to
long-term clinical outcome in the same STEMI population (Paper II)
• To study markers of thrombin generation and their possible association with
myocardial injury and function in the above mentioned STEMI population (Paper
III)
• To study whether circulating levels of CTGF/CCN2 are associated with infarct
size, LV function and remodelling in another cohort of 272 STEMI patients (Paper
IV)
• To study any association between CCN2 and clinical outcome in the two different
STEMI populations mentioned above (Paper IV)
21
Material and methods
The papers in the thesis are based on data from two different study cohorts. All patients have
given written informed consent and the Regional Committee for Medical Research Ethics
approved the studies. Both studies are conducted in accordance with the ethical principles of
the Declaration of Helsinki.
Study subjects and design
BAMI: In the Biobanking in Acute Myocardial Infarction (BAMI) cohort (n=1028), patients
meeting the criteria of an acute STEMI were included the following morning after admission
at Oslo University Hospital, Ullevaal, in the period of June 2007 to August 2011. They were
all investigated with coronary angiography and treated according to international guidelines11.
The exclusion criteria were age <18 years, unstable patient or patient unable to give informed
consent. A blood sample was collected at inclusion, after an overnight fast, median 24h after
start of symptoms and 18h after coronary angiography and, if indicated, percutaneous
coronary intervention (PCI). The samples were stored as the “BAMI biobank” at -80° C until
later analyses of circulating biomarkers from the IL-6 family (Papers I and II), the thrombotic
markers (Paper III) and CCN2/CTGF (Paper IV).
POSTEMI: Briefly, in the Postconditioning in ST-elevation Myocardial Infarction
(POSTEMI) cohort, 272 haemodynamically stable patients with first time STEMI and
symptoms duration < 6 hours were randomised to one of two different reperfusion strategies,
ischemic postconditioning (iPost) or control75. Patients with renal failure (serum creatinine >
200 µmol/L), previous MI, contraindications for CMR or inability to sign informed consent
were excluded. The inclusion period was between January 2009 and August 2012. Blood
samples were drawn before and immediately after the PCI procedure, at Day 1 and after 4
22
months, as described in detail in Paper IV. The primary endpoint of the study was infarct size
measured after 4 months by CMR imaging. Secondary endpoints included myocardial
salvage, microvascular obstruction and clinical events during follow-up.
Coronary angiography and PCI
Coronary angiography was performed in all patients included in the two cohorts. PCI was
performed and adjunctive medical treatment was given according to the present guidelines 11.
Ischemic postconditioning (iPost)
Preconditioning the heart by inducing short episodes of ischemia before a prolonged period of
ischemia has been shown to reduce myocardial injury76. In the setting of an acute STEMI,
preconditioning is not possible and iPost may be an alternative approach. iPost is performed
by different protocols sharing the principle of applying short periods of occlusion of the
infarct related artery during PCI after initial opening of the occluded artery. iPost has been
shown in experimental models and small clinical studies to reduces infarct size76, 77, improve
coronary blood flow and ST-resolution78 and reduce markers of myocardial necrosis (CK
release, troponin)77. The primary objective of the POSTEMI trial was to test whether iPost
could reduce infarct size in STEMI patients76. Briefly, after initial reperfusion with wire and
balloon inflation, the balloon was reinflated 4 times for 1 minute (occluding the vessel),
separated by 1 minute of reflow, and further continuing the procedure after the operators’
judgement (Figure 5). The control group was treated according to standard PCI procedure79.
iPost had no significant effect on infarct size or on secondary ischaemic endpoints in the
study75. Consequently, the POSTEMI population was analysed as a whole in the present
investigation.
23
Figure 5. Flow chart describing the investigational procedure in the POSTEMI trial. Reproduced from Limalanathan with permission. Copyright Clearance Center’s RightsLink® service79.
Blood sampling and laboratory analyses
Detailed descriptions are given in each paper. In the BAMI cohort serum was prepared for
determination of the circulating cytokines and CCN2, and citrated plasma for the haemostatic
variables. Commercially available enzyme-linked immunosorbent assays (ELISA) were used
for all analyses, except for the Calibrated Automated Thrombogram (CAT) assay,
determining the endogenous thrombin potential (ETP), which was used according to the
manufacturer’s instructions (Thrombinoscope BV, Maastricht, The Netherlands) and
measured on the Fluoroscan Ascent fluorometer (Thermo Fisher Scientific OY, Vantaa,
Finland). More detailed description is given in Paper III.
Troponin T, NT-proBNP and other routine blood samples were analysed with conventional
methods, also described in each paper. The coefficients of variation (CVs) for all variables are
described in each paper.
24
Echocardiography
Left ventricular ejection fraction (LVEF) was determined by echocardiography in the BAMI
study (Paper I-III), a.m. Simpson or by visual judgment before hospital discharge or until 3
months after the index event. By several measurements, an average value was estimated.
LVEF data obtained in rehospitalised patients between index event and 3 months later were
not used.
Cardiac magnetic resonance (CMR)
CMR imaging was performed in the acute phase at a median time of 2 days after the index
infarction and repeated at the 4-months’ follow-up visit in the POSTEMI study (Paper IV).
Briefly, a 1.5 T scanner (Philips Intera, release 11 or Philips Achieva, release 3.2, Best,
Netherlands) was used for imaging and further analyses performed on an extended MR Work
Space (Philips Medical Systems). For volume analyses including LVEF, short axis images of
LV were acquired. To analyse the area at risk, defined by a signal intensity (SI) of more than
2 standard deviations above the SI in remote non- infarcted myocardium, T2 weighted
imaging was performed in the short axis plan. For analysing the infarct size, late gadolinium
enhancement (LGE) imaging was used 15 min after contrast injection (0.15 mmol/kg,
gadolinium-DTPA 469 mg/ml (Magnevist; Schering AG, Berlin, Germany) and 2- and 4-
chamber long axis views and short axis views were analysed. Furthermore, myocardial
salvage index was calculated and microvascular obstruction (MVO) was defined as the dark
area within the hyperintense area in the infarcted myocardium. The intra-observer reliability
estimated by the intra-class correlation coefficients for myocardium at risk and infarct size has
been reported80 in detail.
25
Endpoint registration
In the BAMI cohort 989 patients were followed from inclusion until December 2013, median
time 4.6 years (Papers II and IV). Recording of endpoints is described in the papers. In the
POSTEMI study, 272 patients were followed and endpoints were recorded at 4 and 12
months` follow-up. Mortality data were registered from clinical records until median 5.8 years
after inclusion (Paper IV). The primary endpoint in both cohorts was defined as a composite
of all-cause mortality, myocardial infarction, stroke and unscheduled revascularisation ≥3
months after the index event in both cohorts. All-cause mortality was a secondary endpoint in
both cohorts.
Statistics
The statistical approaches and methods used have been described in detail in each paper.
Briefly, continuous data were presented as mean (with standard deviation (SD)) or median
(25th, 75th percentiles) values and categorical data as proportions. Non-parametric tests were
used throughout due to skewed distribution of most of the outcome variables. For association
analyses Spearman’s rho and group analyses (dichotomised or split into quartiles) performed
by Mann-Whitney U test and Kruskal-Wallis test were used for continuous variables, as
appropriate. Chi-square test was used for categorical variables. Multivariate logistic
regression models were used for associations between prothrombotic markers and peak TnT
and LV impairment. Friedman test followed by Wilcoxon signed rank test were used for
comparing CCN2 levels at different sampling points. Cox regression analyses were used for
associations between circulating biomarkers and clinical endpoints. All analyses were
performed by IBM SPSS Software, version 18. (Paper I-III) and version 23.0 (Paper IV)
(SPSS Inc., Chicago, IL).
26
Summary of results
Paper I
We investigated possible associations between members of the IL-6 transsignalling system
including circulating levels of IL-6, sIL-6R, sgp130 and CRP assessed in the acute phase of
an MI and: 1) known risk factors of CAD 2) myocardial injury and LV impairment in 1028
STEMI patients undergoing immediate coronary angiography and PCI if indicated (the BAMI
cohort). Extensive myocardial necrosis defined as the upper quartile of peak TnT (>7140
ng/L), was associated with elevated levels of IL-6 and CRP (p<0.001, both), whereas there
were no significant associations with the novel biomarkers sIL-6R and sgp130 (Figure 1,
Paper I). In addition, levels of IL-6, sgp130 and CRP were higher in patients with elevated
levels of NT-proBNP (> 122ng/L) (p<0.001, p=0.007 and p<0.001, respectively) and with
associations between sgp130 and both treated diabetes mellitus (p=0.002) and abnormal
glucometabolic parameters (HbA1c p<0.001 and admission glucose p=0.02) were observed
Quartiles of peak TnT ng/L
IL-6
pg
/mL
1 2 3 40
10
20
30
40
Quartiles of peak TnT ng/L
sgp
130
ng/
mL
1 2 3 40
100
200
300
Quartiles of peak TnT ng/L
sIL
-6R
ng
/mL
1 2 3 40
20
40
60
Quartiles of peak TnT ng/L
CR
P m
g/L
1 2 3 40
20
40
60
Figure 1 in Paper I. Associations between IL-6, sgp130, IL-6R and CRP and quartiles of peak TnT in 1028 patients with STEMI. Blood was sampled the following morning after admission, median 18 hours after acute coronary angiography
27
Paper II
In 989 STEMI patients from the BAMI–cohort, we examined possible associations between
sIL-6R, sgp130, IL-6 and CRP, and later clinical events in order to obtain new insights into
the importance of the IL-6/gp130 transsignalling pathway in risk stratification of patients with
STEMI.
Altogether there were 200 primary composite endpoints recorded after 4.6 years, 66 deaths,
61 reinfarctions, 6 strokes, 52 urgent PCI, and 15 readmissions for heart failure. All-cause
mortality was a secondary endpoint, and the number was 82.
We found that high levels of sIL-6R defined as the upper quartile (47.7 ng/mL) were
significantly associated with future cardiovascular events, and also with all cause-mortality in
both univariate and multivariate analyses (adjusted HRs 1.54 (95 % CI 1.08, 2.21) (p=0.02)
and 1.81 (1.04, 3.18) (p=0.04), respectively) (Figure 1 and 2 Paper II). Levels of sgp130 were
also higher in patients suffering a new event, however not statistically significant when
adjusted for covariates. CRP was significantly related to all-cause mortality (p=0.01), whereas
no associations between IL-6 levels and cardiovascular events or all-cause mortality were
observed.
28
Figure 1 and 2 in Paper II. Time-to-event curves in 989 STEMI patients for the highest quartile of sIL-6R (Q4) vs. the 3 lowest quartiles (Q1-Q3) according to a composite of cardiovascular events (upper panel) or all-cause mortality (lower panel). The log rank test was used to compare the survival curves.
Paper III
We investigated markers of hypercoagulability in 987 of the STEMI patients from the BAMI
cohort and their associations with myocardial necrosis defined as peak TnT and with LV
function assessed by NT-pro-BNP levels and LVEF. Patients on anticoagulant treatment were
excluded. D-dimer and F1+2, reflecting in vivo thrombin generation and the endogenous
thrombin potential (ETP), reflecting ex vivo potential to thrombin generation, were
investigated.
29
We found both F1+2 and D-dimer to be significantly associated with peak TnT, also after
adjusting for covariates (p<0.001, both) (Figure 1 Paper III, Panel A). Univariate correlations
showed both F1+2 and D-dimer to be significantly associated also with NT-pro-BNP,
however, only D-dimer remained statistically significant after adjustments (p< 0.001) (Figure
1 Paper III, Panel B). Both markers were inversely, but weekly correlated toLVEF (D-dimer:
p<0.001, F1+2; p=0.013), and patients with LVEF <40 % had higher levels of both markers
(p<0.001 and p=0.016, respectively). An inverse pattern for the ex vivo thrombin generation
assessed by ETP was found for peak TnT, NT-proBNP and LVEF.
Figure 1 in Paper III. D-dimer and F1+2 (medians) in quartiles of peak TnT (Panel A) and NT-ProBNP (Panel B). Panel A *=adjusted for age, sex, BMI, hypertension, time from symptoms to blood sampling, CRP and NT-ProBNP; Panel B *=adjusted for age, sex, BMI, hypertension, time from symptoms to blood sampling and CRP
Paper IV
CCN2 has been linked to cardiac hypertrophy and fibrosis. However, some studies have
suggested cardioprotective properties of CCN2 related to LV remodelling after an AMI. Thus,
we aimed to investigate possible associations between circulating CCN2 levels and infarct
size, LV function, adverse remodelling and clinical outcome in STEMI patients.
0
50
100
150
200
250
300
350
Q1 Q2 Q3 Q40
100
200
300
400
500
600
700
Q1 Q2 Q3 Q4
0
50
100
150
200
250
300
Q1 Q2 Q3 Q40
100
200
300
400
500
600
700
Q1 Q2 Q3 Q4
p for trend<0.001* F1+2(pmol/mL)
D-dimer(ng/mL)
p for trend<0.001*
p for trend=0. 324*p for trend=0.001*F1+2(pmol/mL)
We could not demonstrate any significant associations between levels of CCN2 and the
cardiac biomarkers peak TnT and NT-proBNP. Neither were there any significant
associations with CMR measures of myocardial injury, remodelling or function (LVEF,
changes in LV end-diastolic or end-systolic volumes, myocardial salvage or MVO).
Furthermore, we could not find any association between CCN2 and future clinical events
including all-cause mortality, during short-or long-term follow-up, neither in the POSTEMI
cohort nor in the larger BAMI cohort (Figure 2 Paper IV).
Figure 2 in Paper IV. Event-free survival and overall survival according to quartiles of CCN2 measured median 18 hours after PCI in two STEMI cohorts. Upper panels (BAMI) Lower panels (POSTEMI)
31
Discussion
Methodological considerations
There is a selection of patients in both cohorts included in this thesis. Unstable patients (i.e.
heart failure, cardiac arrest, need of assisted ventilation) were excluded in both studies,
resulting in a selection towards somewhat low-risk populations, even though they all had
STEMI. In the BAMI cohort, some patients were excluded for feasibility reasons due to rapid
referral to community hospitals. In the POSTEMI cohort, patients with previous MI,
symptoms >6 hours, severe renal failure (serum creatinine >200 µmol) or contraindications to
CMR were excluded. In both cohorts, the patients had few complications, they had only
slightly reduced LVEF and the mortality rates were low, 8.3 % in BAMI and 2.3 % in the
POSTEMI cohort, which are somewhat lower compared to other reports81, altogether a low-
risk population. However, the large amount of patients in the BAMI cohort, with 20.2 %
events and the long-term follow-up strengthen our results. Another strength of the BAMI
study is that we achieved follow-up data from 989 (96 %) of the 1028 included patients.
The difference in mortality rate between our cohorts may be explained by the difference in
inclusion criteria. In the BAMI cohort, patients with previous MI and heart failure were
included and they had higher frequency of previous hypertension, diabetes mellitus and
hypercholesterolemia compared to the POSTEMI cohort.
In the BAMI cohort, blood samples were collected only once. Thus, the lack of repeated
samples precluded time-course studies of the measured variables and we may have missed the
peak values and possible transient changes, as previously suggested for some biomarkers82, 83.
In the BAMI cohort, patients with more than 6 hours from symptoms to admission were also
included, thus there was a large variation in time from start of symptoms to blood sampling
32
(2-264 hours). There was a significant correlation between time and CRP (r=0.45, p<0.001),
but only a weak correlation between time and IL-6 (r=0.08, p=0.01). This is in line with other
reports showing peak IL-6 the first day after admission in PCI-treated STEMI patients84, 85,
whereas peak level of CRP seems to be as late as 72 hours after PCI85. This is expected from
the well-described induction of CRP expression in the liver by IL-6, which also fits the strong
correlation we observed between CRP and time. However, adjusting for the time frame did
not change our results. There were no significant correlations between time and the other
measured parameters, suggesting the variables to be representative of a STEMI cohort. Also,
adjusting for the time frame did not change the results described in Paper III.
In the BAMI cohort (Papers I-III), the echocardiography investigations were performed by
different investigators and at different time points, which may represent a potential bias.
In the POSTEMI cohort (Paper IV), the serial measurements of CCN2 in the acute phase of a
STEMI strengthen the results, but the lack of blood samples between the acute phase and 4
months is a limitation and we may speculate whether we missed the peak value, i.e. that
CCN2 peaked at a later time point.
In the POSTEMI cohort, the repeated CMR measurements in a relative large number of
patients strengthen our results on myocardial injury and function. Finally, the CCN2 results
regarding clinical outcome are strengthened by the fact that CCN2 was measured in two
different cohorts, both with long-term clinical follow-up.
33
General discussion
Members of the IL-6 family in patients with acute myocardial infarction
We showed in our study that IL-6 and CRP were significantly associated with myocardial
necrosis and impaired LV function, which confirms previous studies39, 59, 86 and fits with the
assumption of a connection between inflammation and infarct size. IL-6 is a pleiotropic
multifunctional cytokine, and both pro- and anti-inflammatory effects of IL-6 have been
suggested35, 36. The diversity of the IL-6 signalling is complex and not fully understood. As
described in the Introduction chapter, most cells express gp130 on the cell surface, while few
cells have membrane-bound IL-6 receptors35. Only cells that have IL-6 receptors on their cell
membrane can bind IL-6 and induce the classical signalling, while the complex of IL-6 and
sIL-6R can bind to all cells that express gp130, inducing the transsignalling. The soluble form
of IL-6R will therefore have a much broader spectrum of target cells35. Transsignalling
mediated by increased levels of soluble IL-6R seems to be responsible for activation of the
immune system and recruitment of mononuclear cells seen in chronic inflammatory diseases
like Crohn`s disease and rheumatoid arthritis35, 55 and possibly the atherosclerotic process.
The classical pathway is also thought to play a role in the early immune response by induction
of the acute-phase IL-6, in which levels in the physiological range are necessary for the
inflammatory response36. However, an excessive activation in the acute phase or the on-going
chronic activation may have detrimental effects by its action trough the transsignalling
pathway35, 36, 55. The exact roles of IL-6, sIL-6R and the signalling receptor protein gp130 are,
however, not fully understood and the effects of inhibition of the different members of the IL-