Coronary microcirculation and peripheral endothelial function evaluation after acute ST elevation myocardial infarction treated with primary angioplasty

CORONARY MICROCIRCULATION AND PERIPHERAL ENDOTHELIAL FUNCTION EVALUATION AFTER ACUTE ST ELEVATION MYOCARDIAL INFARCTION TREATED WITH PRIMARY ANGIOPLASTY SÉRGIO BRAVO CORDEIRO BAPTISTA Thesis for the PhD degree in Medicine Subspecialty of Clinical Investigation NOVA Medical School, Lisbon September 2016 CORONARY MICROCIRCULATION AND PERIPHERAL ENDOTHELIAL FUNCTION EVALUATION AFTER ACUTE ST ELEVATION MYOCARDIAL INFARCTION TREATED WITH PRIMARY ANGIOPLASTY SÉRGIO BRAVO CORDEIRO BAPTISTA Supervisor: Professor Doutor Victor Gil Co-supervisor: Professor Doutor José Fragata Thesis for obtaining the PhD degree in Medicine, Subspecialty Clinical Investigation September 2016 MICROCIRCULAÇÃO CORONÁRIA E A AVALIAÇÃO DA FUNÇÃO ENDOTELIAL PERIFÉRICA APÓS ENFARTE AGUDO DO MIOCÁRDIO COM ELEVAÇÃO DO ST TRATADO POR INTERVENÇÃO CORONÁRIA PERCUTÂNEA PRIMÁRIA SÉRGIO BRAVO CORDEIRO BAPTISTA Tese para a obtenção do grau de Doutor em Medicina na especialidade de Investigação Clínica na Faculdade de Ciências Médicas de Lisboa Setembro de 2016 MICROCIRCULAÇÃO CORONÁRIA E A AVALIAÇÃO DA FUNÇÃO ENDOTELIAL PERIFÉRICA APÓS ENFARTE AGUDO DO MIOCÁRDIO COM ELEVAÇÃO DO ST TRATADO POR INTERVENÇÃO CORONÁRIA PERCUTÂNEA PRIMÁRIA SÉRGIO BRAVO CORDEIRO BAPTISTA Orientador: Professor Doutor Victor Gil Coorientador: Professor Doutor José Fragata Tese para a obtenção do Grau de Doutor em Medicina na especialidade de Investigação Clínica Setembro de 2016 Para a Teresa, e para a Mariana, a Sofia e o Miguel Contents 16 1.4. Secondary endpoint: Extent of myocardial infarction and microvascular reperfusion according to RHI values on the second EndoPAT ................................... 166 1.5. Secondary endpoint: Extent of myocardial infarction and microvascular reperfusion according to IMR values ......................................................................... 167 DISCUSSION ....................................................................................................................... 169 1. Main findings of the study ................................................................................................. 171 2. Patient inclusion and validity of the sample ...................................................................... 172 3. First EndoPAT results ......................................................................................................... 174 3.1. Implications of the RHI results in the first EndoPAT evaluation ................................ 174 3.2. IMR, microvascular obstruction and extent of infarction according to RHI on the first EndoPAT .............................................................................................................. 178 4. Second EndoPAT results .................................................................................................... 179 4.1. RHI results in the second EndoPAT evaluation .......................................................... 179 4.2. IMR, microvascular obstruction and extent of infarction according to RHI on the second EndoPAT ........................................................................................................ 180 5. IMR results ......................................................................................................................... 182 5.1. IMR and microvascular obstruction ........................................................................... 182 5.2. IMR and infarct extension .......................................................................................... 183 5.3. IMR and left ventricular remodelling ......................................................................... 183 5.4. Predictors of IMR ....................................................................................................... 184 STUDY LIMITATIONS AND STRENGTHS ................................................................................ 187 CONCLUSIONS AND FUTURE PERSPECTIVES ........................................................................ 191 BIBLIOGRAPHY ................................................................................................................... 195 ACKNOWLEDGMENTS ........................................................................................................ 213 APPENDIX – PAPERS ..................................................................... Error! Bookmark not defined. 1 – Endothelial dysfunction evaluated by peripheral arterial tonometry is related with peak TnI values in patients with ST elevation myocardial infarction treated with primary angioplasty ................................................. Error! Bookmark not defined. 2 – The Index of Microcirculatory Resistance as a Predictor of Echocardiographic Left Ventricular Performance Recovery in Patients With ST-Elevation Acute Myocardial Infarction Undergoing Successful Primary Angioplasty ..... Error! Bookmark not defined. 3 – Early peripheral endothelial dysfunction predicts myocardial infarct extension and microvascular obstruction in patients with ST elevation myocardial infarction .......................................................................... Error! Bookmark not defined. Statement of Responsibility 17 Contributions and responsibilities The candidate was responsible for the design, implementation (including logistics and financial), patient inclusion, data registry and statistical analysis of the study, as well as for the full writing of this thesis. All patients included were admitted to Hospital Prof. Doutor Fernando da Fonseca´s Cardiology Department (Director until 2013: Prof. Doutor Victor Gil; since 2013: Dr. Carlos Morais). The acute invasive procedure (IMR measurement) was performed either by the candidate, or by one of the Interventional Cardiologists of the Interventional Cardiology Unit of Hospital Prof. Doutor Fernando da Fonseca (Coordinator: Dr. Pedro Farto e Abreu). The EndoPAT evaluations were performed in the Endothelial Function Laboratory of Hospital Prof. Doutor Fernando da Fonseca (Director: Prof. Doutor Victor Gil), either by the candidate, Dra. Mariana Faustino or one of the trained technicians. The Echocardiographic evaluations were performed in the Echocardiography Unit of Hospital Prof. Doutor Fernando da Fonseca (Coordinator: Dr. António Freitas), by adequately trained physicians. Their analysis was performed by Dr. António Freitas and Dra. Mariana Faustino. The ECGs were performed by trained cardiopulmonary technicians. All laboratory evaluations were performed by the Clinical Pathology Department of Hospital Prof. Doutor Fernando da Fonseca (Director: Dra. Luísa Sancho). The cardiac magnetic resonance exams were performed in Hospital dos Lusíadas, Lisbon (Cardiovascular Department, coordinated by Prof. Doutor Victor Gil and Imaging Department, coordinated by Drª Graça Correia), by Dr. João Abecasis, who was also responsible for their interpretation. Publications 19 LIST OF PUBLICATIONS Papers in Periodicals with Referees This thesis resulted so far in the following publications:  Baptista SB, Faustino M, Simões J, Nédio M, Monteiro C, Lourenço E, Leal P, Farto e Abreu P, Gil V. Endothelial dysfunction evaluated by peripheral arterial tonometry is related with peak TnI values in patients with ST elevation myocardial infarction treated with primary angioplasty. Microvasc Res. 2016 May;105:34-9. doi: 10.1016/j.mvr.2015.12.010. Epub 2015 Dec 22.  Faustino M, Baptista SB, Freitas A, Monteiro C, Leal P, Nédio M, Antunes C, Abreu PF, Gil V, Morais C. The Index of Microcirculatory Resistance as a Predictor of Echocardiographic Left Ventricular Performance Recovery in Patients With ST-Elevation Acute Myocardial Infarction Undergoing Successful Primary Angioplasty. J Interv Cardiol. 2016 Apr;29(2):137-45. doi: 10.1111/joic.12278. Epub 2016 Mar 1.  Baptista SB, Faustino M, Brizida L, Loureiro J, Augusto J, Abecasis J, Monteiro C, Leal P, Nédio M, Farto e Abreu P, Gil V, Morais C. Early peripheral endothelial dysfunction predicts myocardial infarct extension and microvascular obstruction in patients with ST elevation myocardial infarction. Submitted A forth manuscript is currently being prepared:  Baptista SB, Faustino M, Loureiro J, Brízida L, Freitas A, Augusto A, Abecasis J, Monteiro C, Leal P, Nédio M, Antunes C, Farto e Abreu P, Gil V, Morais C. The index of microvascular resistance as a surrogate for myocardial infarct extension, microvascular obstruction and left ventricular remodelling in patients with ST elevation myocardial infarction treated by primary angioplasty. Papers in Conference Proceedings This thesis resulted in the following publications as abstracts in conference proceedings:  Baptista SB, Faustino M, Gil V. Severity Of Coronary Artery Disease And Endothelial Dysfunction Evaluated By Peripheral Arterial Tonometry. Journal of the American College of Cardiology 2013;62 (18_S1):B183-B183 Publications 20  Baptista SB, Faustino M, Abreu PF, Gil V. Impact of endothelial dysfunction evaluated by peripheral arterial tonometry in the extension of ST elevation myocardial infarction treated with primary angioplasty. Journal of the American College of Cardiology 2013;62 (18_S1), B183-B183  Baptista SB., Faustino M., Loureiro J., Brizida L., Alves P., Nedio M., Monteiro C., Lourenço E., Leal P., Farto e Abreu P., Gil V., Morais C. Index of microcirculatory resistance in patients with STEMI treated with primary angioplasty: relation with other indicators of microvascular reperfusion, infarct size and regional systolic function. Eurointervention 2015; Abstracts EuroPCR 2015; Euro15A-MA006.  Baptista SB., Faustino M., Loureiro J., Brizida L., Alves P., Nedio M., Monteiro C., Lourenço E., Leal P., Farto E Abreu P., Gil V., Morais C. Predictors of microcirculatory dysfunction in STEMI patients treated with primary angioplasty. Eurointervention 2015; Abstracts EuroPCR 2015; Euro15A-MA009  Faustino M, Baptista SB, Freitas A, Monteiro C, Leal P, Nédio M, Antunes C, Farto e Abreu P, Gil V, Morais C. Coronary index of microcirculatory resistance and echocardiographic parameters evolution in patients with ST-elevation acute myocardial infarction treated with primary angioplasty. European Heart Journal 2015;36 (suppl 1):240  Faustino M, Baptista SB, Freitas A, Augusto JA, Leal P, Nédio M, Antunes C, Abreu PF, Gil V, Morais C. Global longitudinal strain and coronary microcirculation status in prediction of left ventricle functional recovery after ST elevation myocardium infarction. Eur Heart J Cardiovasc Imaging 2015;16(suppl 2):S156-S182. List of Abbreviations 21 LIST OF ABBREVIATIONS ∆P Pressure gradient 2D Two-dimensional ACE Angiotensin-converting-enzyme ACEi Angiotensin-converting-enzyme inhibitors ACS Acute coronary syndromes ADMA Asymmetrical dimethylarginine ARB Angiotensin II receptor blockers AUC Area under the curve BMI Body mass index BNP B-type natriuretic peptide CABG Coronary artery bypass graft CAD Coronary artery disease CAMs Cellular adhesion molecules ceCMR Contrast enhanced cardiac magnetic resonance CFR Coronary flow reserve CI Confidence interval CK Creatine kinase CKD Chronic kidney disease CK-Mb Mb fraction of creatine kinase CT Cardiac tomography cTFC Corrected TIMI frame count CV Cardiovascular DBP Diastolic blood pressure DICOM Digital Imaging and Communications in Medicine ECG Electrocardiogram Echo Echocardiogram EDHF Endothelium-dependent hyperpolarizing factor EECP Enhanced external counterpulsation FMD Flow-mediated dilation FRS Framingham Risk Score GLS Global longitudinal strain HbA1c Glycosylated haemoglobin HF Heart failure hMR Hyperaemic microvascular resistance index HR Hazard ratio hs-CRP High sensitive C-reactive protein IMR Index of microcirculatory resistance IMRtrue Index of microcirculatory resistance corrected for collateral flow IQR Interquartile range IS Infarct size ISR In-stent restenosis Publications 22 IVUS Intravascular ultrasound L_RHI Logarithmic transformation of reactive hyperaemia index LA Left atrium LAD Left anterior descending artery LCx Left circumflex artery LDF Laser Doppler flowmetry LDL-C Low-density lipoprotein cholesterol L-FMC Low flow mediated constriction Lp-PLA2 Lipoprotein-associated phospholipase A2 LV Left ventricle LVEdV Left ventricular end diastolic volume LVEF Left ventricular ejection fraction LVEsV Left ventricular end systolic volume MACE Major adverse cardiac events MBG Myocardial blush grade MCE Myocardial contrast echocardiography MI Myocardial infarction MVO Microvascular obstruction NO Nitric oxide NST-ACS Non-ST elevation acute coronary syndrome NSTEMI Non-ST elevation myocardial infarction NT-pro-BNP N-terminal pro b-type natriuretic peptide NYHA New York Heart Association classification for heart failures symptoms OR Odds ratio OxLDL Oxidized low-density lipoprotein Pa Aortic pressure PAI-1 Plasminogen activator inhibitor-1 PAT Peripheral arterial tonometry PCI Percutaneous coronary intervention Pd Distal pressure PET Positron emission tomography PGI2 Prostaglandin I2 (Prostacyclin) P-PCI Primary (urgent) percutaneous coronary intervention PzF Zero-flow pressure Q Flow R Resistance RCA Right coronary artery RHI Reactive hyperaemia index ROC Receiver operating characteristic SBP Systolic blood pressure SK Streptokinase SMC Smooth muscle cells SPECT Single-photon emission computed tomography STEMI ST elevation myocardial infarction List of Abbreviations 23 SYNTAXsc Synergy Between PCI With Taxus and Cardiac Surgery Score Tmn Mean transit time TMPG TIMI myocardial perfusion grade TnI Troponin I TnIAUC Area under the curve of troponin TnIpeak Peak value of troponin t-PA Tissue plasminogen activator VOP Venous occlusion plethysmography WMSI Wall motion score index LIST OF STUDY ACRONYMS APPROACH Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease ARTS Arterial Revascularization Therapies Study BARI Bypass Angioplasty Revascularization Investigation Myocardial Jeopardy PROSPECT Providing Regional Observations to Study Predictors of Events in the Coronary Tree SYNTAX Synergy Between PCI With Taxus and Cardiac Surgery TIMI Thrombolysis In Myocardial Infarction List of figures 25 LIST OF FIGURES Figure 1 – Coronary arterial circulation .......................................................................................... 38 Figure 2 – Schematic of physiological assessment (FFR, CFR and IMR) using a coronary pressure and temperature wire ..................................................................................................... 53 Figure 3 – Reactive hyperemia index according to the severity of the coronary artery disease (number of major vessels with lesions >70%) ................................................................... 73 Figure 4 – Peak Troponin I values according to the presence of endothelial dysfunction (RHI<1.67) in patients with ST elevation myocardial infarction treated with primary PCI ............ 75 Figure 5 – Modified APPROACH score for evaluating the area-at-risk ........................................... 87 Figure 6 – Area-at-risk calculation according to the APPROACH and BARI scores ......................... 88 Figure 7 – The Certus Pressure-wire, with a pressure and temperature sensor, and the RadiAnalyzer Xpress (St. Jude Medical®) ........................................................................................ 91 Figure 8 – Example of baseline and hyperaemic thermodilution curves ....................................... 92 Figure 9 – The EndoPAT 2000 ......................................................................................................... 95 Figure 10 – EndoPAT procedure ..................................................................................................... 96 Figure 11 - Flow diagram of the cohort study .............................................................................. 101 Figure 12 – Age distribution of patients included ........................................................................ 102 Figure 13 – Prevalence of main risk factors ................................................................................. 102 Figure 14 – Correlation between APPROACH and BARI scores in identifying the area-at-risk for each lesion location ................................................................................................................ 103 Figure 15 – Corrected TIMI frame count, according to TIMI myocardial perfusion grade .......... 104 Figure 16 – Examples of IMR measurement ................................................................................ 105 Figure 17 – Reactive hyperaemia index (RHI) values on the 1st EndoPAT (complete cohort) ..... 107 Figure 18 – Reactive hyperaemia index (RHI) values on the 1st EndoPAT (excluding outliers) .... 107 Figure 19 - Examples of peripheral endothelial function measured by digital pulse amplitude with endothelial peripheral arterial tonometry.......................................................... 109 Figure 20 – Three examples of ceCMR ......................................................................................... 113 Figure 21 – Boxplot of IMR values according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation ................................................................................. 119 Figure 22 – Boxplot of IMR values according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation ............................................................................ 129 Figure 23 – ST elevation and deviation resolution (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation ...................... 140 Figure 24 – ST residual changes (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation ............................................................. 141 Figure 25 – ST elevation and deviation resolution (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation ................ 148 Figure 26 – ST residual changes (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation ........................................................ 149 Figure 27 – ST elevation and deviation resolution (median values) according to the presence of coronary microvascular dysfunction (IMR>24) ........................................................ 159 Resumo 32 RESUMO Introdução: Apesar da normalização do fluxo coronário epicárdico após intervenção coronária percutânea primária (ICP-P), uma proporção significativa dos doentes com enfarte agudo do miocárdio com elevação do segmento ST (EAMcST) têm piores resultados clínicos devido ao desenvolvimento de lesão ou disfunção microvascular coronária. A disfunção endotelial provavelmente desempenha um papel nesta lesão microvascular coronária e a sua avaliação por tonometria arterial periférica poderá ser útil para prever a extensão da lesão microvascular e a extensão do enfarte. Objectivos: Avaliar a relação da disfunção endotelial periférica precoce, avaliada pelo índice de hiperémia reactiva (IHR, obtido por tonometria arterial periférica) com o índice de resistência da microcirculação (IRM), medido imediatamente após a ICP-P e estimar a relação entre o IHR e o IRM e, 1) a extensão do enfarte, avaliada por ressonância magnética cardíaca com contraste (RMCc) e pela curva de libertação de Troponina I; 2) a extensão da obstrução microvascular (OMV), avaliada por RMCc e por outros indicadores indirectos; 3) a remodelagem ventricular esquerda tardia (aos 3 meses), avaliada por ecocardiografia. Métodos. Estudo observacional, prospectivo, de coorte. Foram incluídos doentes com um primeiro EAMcST, tratados com sucesso por ICP-P, hemodinamicamente estáveis e sem contraindicações para administração de adenosina. Depois da ICP-P, o IRM foi medido usando um fio de pressão. O IHR foi avaliado na fase aguda e novamente 24 horas depois da ICP-P. A disfunção endotelial foi definida como um IHR<1,67 e o IHR foi também analisado por tercis. Os indicadores angiográficos de reperfusão (contagem corrigida de frames e grau de perfusão miocárdica TIMI) foram avaliados no final da ICP-P. Foram colhidas análises na admissão e em horários definidos nas primeiras 48 horas para avaliação da Troponina I. Antes, imediatamente após e 90 e 180 minutos depois da ICP-P foram registados electrocardiogramas, para avaliação da resolução das alterações do segmento ST. A função ventricular esquerda global e segmentar foi avaliada por ecocardiografia após a ICP-P e aos 3 meses. A RCMc foi efectuada ao 7-8º dia após o EAMcST. Resultados: Foram incluídos 60 doentes (48 homens, idade media 59,6±12,7 anos). Na primeira avaliação, os valores de IHR foram muito superiores ao esperado (média 2,15±0,58), provavelmente por erros técnicos incontornáveis, não se relacionando com nenhum dos indicadores de extensão do enfarte ou de OMV. Na segunda avaliação, às 24h, os valores médios de IRH foram 1,87±0,60. Os doentes com IRH <1,67 tiveram tendencialmente valores mais elevados de IRM (mediana 40,5 IIQ 54,4 vs. mediana 22,0 IIQ 26,0, p=0,09), pior resolução do segmento ST, piores resultados nos indicadores angiográficos de OMV e maior probabilidade de ter OMV na RMNc (54,1% vs. 11,1%, p=0,03). Também tiveram enfartes de maior dimensão na avaliação pela TnI I máxima (p=0,004) e pela área sob a curva de TnI (p= 0,012). A fracção de ejecção do ventrículo esquerdo (FEVE) foi menor e o score de motilidade segmentar (SMS) maior nestes doentes. A mediana do IRM foi 24 (IIQ 33). O IRM correlacionou-se fortemente com a OMV avaliada na RMNc (r=0.91, p<0.001; curva ROC 0,723, IC95% 0,500-0,896, p=0,018). Nos doentes com IRM >24, a resolução do ST foi significativamente menor e os indicadores angiográficos de reperfusão foram significativamente piores. O IRM também se correlacionou com a massa de enfarte (r=0,70, p<0,001) e a massa de miocárdio salvo (r=0,35, p=0,014) na RMCc. Os doentes com IRM>24 tiveram valores significativamente mais elevados de TnI máxima (p=0,013) e ASC de TnI (p=0,003). A FEVE melhorou de forma significativa apenas nos doentes com IMR<24 (p=0,01). Os preditores independentes do IRH foram a idade, a glicemia na admissão e a HbA1c na admissão. Conclusões: Não parece ser possível avaliar de forma fidedigna o IHR na fase aguda do EAMcST após ICP-P. O IHR medido 24h após a ICP-P é mensurável de forma adequada e prevê a dimensão do enfarte e da OMV, confirmando a disfunção endotelial como um mecanismo importante na disfunção microvascular em doentes com EAMcST. O IRM correlaciona-se fortemente com a OMV e permite prever a dimensão do enfarte e o risco de remodelagem ventricular esquerda. 33 INTRODUCTION Introduction 35 1. The scope of the problem: microcirculatory dysfunction in patients with acute STEMI The interventional cardiologists that perform primary percutaneous coronary interventions have a common insecurity when they face a patient with an occluded coronary artery in the setting of an acute ST elevation myocardial infarction: what if opening the artery is not enough? This anguish can take several forms: either more immediate – like the no-reflow phenomenon, the inability to improve patient’s pain or the unexpected absence of ST resolution – or later – like unexpected severe left ventricular remodelling or left ventricular dysfunction in the follow-up, even after a successful and timely procedure. All these questions and fears seem to have a common ground, one we need to understand better in order to further improve STEMI patients outcomes: the coronary microcirculation. The diagnosis and treatment of acute ST elevation myocardial infarction (STEMI) has undoubtedly been the subject of intense investigation over the last decades. As a result, prompt implementation of both mechanical (primary percutaneous coronary intervention) and pharmacological (adjuvant anti-platelets and anticoagulants) approaches to reopen the occluded coronary artery is nowadays universally accepted as the treatment of choice to improve survival in STEMI patients.1,2 Notwithstanding all the improvements we have seen in primary PCI programs, with dramatic reductions in the time between symptom onset and the intervention, and despite a normal epicardial coronary artery flow is almost always achieved in a timely fashion after primary percutaneous coronary intervention (P-PCI), a significant proportion of patients (from 20% to 60%) has a poor outcome because of microvascular coronary damage.3,4 In fact, microvascular perfusion is often impaired after P-PCI, and reperfusion of the epicardial coronary artery will not always guarantee reperfusion at the myocardial tissue level5. The extent of this microvascular coronary dysfunction has been shown to be an important and independent contributor to subsequent changes in left ventricular geometry and performance.6,7 Patients with impaired microvascular perfusion have larger infarcts, as evaluated by CK and troponin release, less electrocardiographic ST elevation resolution, larger long-term left ventricular wall motion abnormalities and lower left ventricular ejection fraction, and larger necrotic areas as evaluated by cardiac magnetic resonance. As a consequence of these, coronary microvascular dysfunction is associated with higher event rates, risk of progression to heart failure and mortality. The precise mechanisms underlying coronary microcirculation dysfunction before and after the restoration of epicardial blood flow are largely unknown and likely to be multifactorial. Understanding what happens in the microcirculation in the setting of a ST elevation myocardial infarction is thus very relevant. Most research done on this matter, however, has focused on the changes occurring on microcirculation after the STEMI and very little is known about the effect of pre-existent microvascular coronary dysfunction. Introduction 36 The recent development of invasive techniques for the precise evaluation of the coronary microcirculation (such as the resistance – IMR) and of non-invasive methods for the evaluation of endothelial dysfunction (such as digital peripheral arterial tonometry - PAT) opens a new window of opportunity for the understanding of the pathophysiological processes occurring in coronary microvasculature in patients with STEMI. I have been deeply involved in the area of functional evaluation of coronary disease in the past few years, both locally8–13and at a national14 and international levels15–17 and I am currently involved in several trials evaluating new non-hyperaemic technologies for physiological assessment of intermediate lesions and/or new clinical indications. Therefore, both I and all the Interventional Cardiology Department in which I work are very familiar with the use of pressurewire technology, both for clinical and investigational purposes. On the other hand, I am also quite familiar with the PAT technology, following the creation of an Endothelial Function Laboratory by Prof. Victor Gil. This laboratory has conducted several studies, not only in coronary artery disease,18–21, but also in other clinical conditions, like rheumatologic diseases22, obesity23–25 or erectile dysfunction26, in which I had the opportunity to collaborate. Therefore, using these new invasive (IMR) and non-invasive (EndoPAT®) techniques with the intention to improve our understanding of the coronary microcirculation dysfunction in STEMI patients was a natural step. In the following pages, the evidence that supports this study will be presented, organized in six sections:  Section 2 is a short description of the normal coronary microcirculation.  Section 3 presents the changes thought to occur in acute STEMI and the prevailing theories on the role of microcirculation, with particular focus on the endothelial function and on the gaps in knowledge that led to the current study.  Section 4 outlines the tools available for evaluating coronary microvascular circulation.  Section 5 focus on the index of microcirculatory resistance the clinical evidence currently available.  Section 6 is dedicated to describing the tools available for endothelial function evaluation.  Section 7 further details non-invasive peripheral arterial tonometry (EndoPAT). Finally, a short summary of the fundaments of this thesis are presented in Section 8. Introduction 37 2. The normal coronary microcirculation 2.1. Anatomy and function of the coronary arterial system The coronary arterial system can be subdivided into three functional compartments: conductive vessels, pre-arteriolar vessels and arterioles.27–29 The coronary blood flow is driven by the pressure difference between the aortic sinus and the coronary sinus (or the right atrium pressure) across these three compartments:  Conductive vessels (corresponding to epicardial arteries) are the first compartment and in the absence of obstructive stenosis, they offer very little resistance to coronary blood flow (even at maximum hyperemia), serving mainly as conductance vessels. Approximately 60% of their wall thickness consists of the muscular media, which can respond to changes in aortic pressure and modulates coronary tone in response to flowmediated endothelium-dependent vasodilators, circulating vasoactive substances and neural stimuli.  The intermediate compartment is represented by pre-arterioles, which are resistive vessels connecting the conductive arteries to the arterioles. They are epicardial (extramyocardial) vessels and have a diameter in the range of 200 to 500 μm. These vessels react to changes in shear stress and intravascular pressure to preserve adequate perfusion pressure in the distal arteriolar bed, being responsible for approximately 25% of the total coronary vascular resistance.  The distal compartment consists of arterioles. They are smaller than 200 μm in diameter and are the main regulatory component of the coronary circulation, representing approximately 55% of the total coronary vascular resistance. Arterioles are usually subdivided in two categories, according to their diameter and the mechanism(s) that regulate their tone30: endothelium-dependent vasoreactivity prevails in the larger arterioles (100–200 μm) and translates flow-related stimuli into vasomotor responses, i.e. vasodilation with increase in flow and vice versa. Medium-sized microvessels (40–100 μm in diameter) react predominantly to intraluminal pressure changes sensed by stretch receptors located in vascular smooth muscle cells (myogenic control), i.e. they constrict when the intraluminal pressure increases and, conversely, dilate when the pressure decreases. Finally, the tone of the smaller arterioles (vessels less than 40 mm in diameter) is modulated by the metabolic activity of the myocardium. The arterioles are responsible for the process of coronary autoregulation31,32: coronary flow is regulated independently of the arterial perfusion pressure despite large variations in this pressure. Vasodilatation of the smaller arterioles is induced by increased metabolic activity, which leads to pressure reduction in the medium-sized microvessels and myogenic dilation, which, in turn, increases flow upstream resulting in endothelium-dependent vasodilation (Figure 1). Introduction 38 Figure 1 – Coronary arterial circulation Conductive vessels and proximal prearterioles are most responsive to flowdependent dilation. Distal prearterioles are most responsive to changes in intravascular pressure (and are mainly responsible for autoregulation of coronary blood flow), whereas arterioles are most responsive to changes in the myocardial concentration of metabolites, and are mainly responsible for the metabolic regulation of coronary blood flow (Adapted from Camici et al33) These mechanisms effectively and efficiently allow the microcirculation to regulate myocardial perfusion both at rest and at different levels of myocardial metabolic demand.30 When pressure falls to the lower limit of autoregulation, coronary resistance arteries are maximally vasodilated by intrinsic stimuli, and flow becomes pressure-dependent, resulting in the onset of subendocardial ischemia. Resting coronary blood flow under normal hemodynamic conditions averages 0.7 to 1.0 mL/min/g and can increase four to fivefold during vasodilation.34 2.2. Regulation of coronary vascular tone Coronary blood flow is adjusted to the metabolic needs of the myocardium by at least three essential regulators of coronary tone:  The metabolic vasodilatory system, through the production of adenosine and simultaneous opening of the ATP-dependent K+ channels within the myocardial cells (both potent vasodilators) in the presence of an increase in the oxygen consumption and metabolic demand35;  The neurogenic control system, through adrenergic activation of α-receptors, resulting in arteriolar vasoconstriction. Vasoconstriction mediated by α1-receptor acts mainly on the larger coronary arteries, whereas both α1and α2-receptor activity is involved in Introduction 39 regulating the degree of vasoconstriction of the smaller resistance vessels.29 These influences are opposed by the vasodilatory effect of vascular β-receptor stimulation and metabolic mechanisms. Importantly, cholinergic stimulation, normally vasodilatory because it releases nitric oxide, becomes vasoconstrictive when the endothelium is damaged (see below).  The vascular endothelium, that acts both through vasodilation and vasoconstriction, playing a central role in the regulation of vascular tone, since it closely interacts with the two other systems involved: o Endothelial-dependent vasodilation is mediated by nitric oxide (which is a very short-lived vasodilating factor) and by the endothelium-dependent hyperpolarizing factor (EDHF), both secreted by healthy endothelial cells in response to i) vascular shear forces associated with increased coronary flow36; ii) adenosine (acting on endothelial ATP-dependent K+ channels37); iii) or other agonists (e.g. released from platelets or cardiac nerves). Prostacyclin, or prostaglandin I2 (PGI2), is also produced in the coronary endothelium of collateral vessels and causes tonic vasodilation. o Endothelial-dependent vasoconstriction is mediated by endothelin-1 which activates protein kinase C in vascular smooth muscle to produce coronary constriction and competes with endothelium-derived relaxing factors. This effect is enhanced in diseased atherosclerotic arteries with extensively damaged endothelium.38 2.3. The vascular endothelium The vascular endothelium is a monolayer of cells covering the internal lumen of all blood vessels, separating the blood from the vascular wall and organ tissues. The vascular endothelium has different functions39:  It plays a crucial role in vascular tone and blood flow regulation (as described in the previous section);  It is an anti-coagulant surface, with an active role on platelet adhesion and aggregation, and on thrombosis.40 Under physiological conditions, the endothelium prevents thrombus formation through a number of mechanisms: thrombomodulin, protein S, heparin sulphate, and tissue factor pathway inhibitor are all endothelium-derived inhibitors of coagulation, whereas PGI2, NO, and surface-bound CD39 inhibit platelet aggregation.41  It regulates vascular permeability to plasma constituents between blood and tissues42;  It contributes to vascular homeostasis and repair.43 Additionally, the endothelium also actively produces proinflammatory and anti-inflammatory molecules, and is the actual target for circulating inflammatory mediators that are synthesized by other cell types, including platelets, leukocytes, hepatocytes, and adipocytes.41 Introduction 40 3. Coronary microcirculation dysfunction 3.1. Classification and mechanisms of coronary microcirculation dysfunction Coronary microvascular dysfunction has been widely studied in the last two decades, particularly in patients chest pain despite having normal coronary arteriograms (i.e. microvascular angina). In an attempt to summarize the mechanisms involved and its clinical translation, Camici and Crea proposed44 (and recently reviewed45) a classification for the different clinical types of coronary microvascular dysfunction (Table 1). Of the four types of microvascular dysfunction proposed, Types 3 (obstructive coronary artery disease) and 4 (iatrogenic) are the ones involved in patients with acute ST elevation myocardial infarction. The authors also put forward the main pathogenic mechanisms more likely to be involved in these types of coronary microvascular dysfunction (luminal obstruction by thrombotic debris, endothelial dysfunction, smooth muscular cell dysfunction and autonomic dysfunction). Table 1 – Classification of coronary microvascular dysfunction Clinical Setting Main pathogenic mechanisms Type 1: In the absence of myocardial diseases and obstructive CAD Risk factors Microvascular angina Endothelial dysfunction SMC dysfunction Vascular remodelling Type 2: In myocardial diseases Hypertrophic cardiomyopathy Dilated cardiomyopathy Anderson-Fabry’s disease Amyloidosis Myocarditis Aortic stenosis Vascular remodelling SMC dysfunction Extramural compression Luminal obstruction Type 3: In obstructive CAD Stable angina Acute coronary syndrome Endothelial dysfunction SMC dysfunction Luminal obstruction Type 4: Iatrogenic PCI Coronary artery grafting Luminal obstruction Autonomic dysfunction Adapted from Camici and Crea.45 CAD: coronary artery disease; SMC: smooth muscle cells; PCI: percutaneous coronary intervention However, the exact role of each of the abovementioned mechanisms is difficult to identify, particularly in pathogenic processes. In fact, a substantial number of hypotheses for the pathophysiology of coronary microvascular dysfunction has been proposed in the literature44,46 including:  Structural changes: o Luminal obstruction, caused by microembolization in acute coronary syndromes or after recanalization. Introduction 41 o Vascular wall infiltration (e.g. in infiltrative diseases, like Anderson-Fabry cardiomyopathy) o Altered microvascular remodelling, through sclerosis of small arteries and arterioles with perivascular fibrosis, swollen endothelial nuclei in capillaries, and irregular lumina of small arteries.47 o Vascular rarefaction and perivascular fibrosis (e.g. in aortic stenosis and arterial hypertension).  Functional changes: o Endothelial dysfunction, including the release of substances with opposing effects (such as endothelin, thromboxane A2, prostaglandin H2, and superoxide) and resulting in a shift from a net dilator response to a net constrictor response to a variety of stimuli. Endothelial dysfunction also involves a switch from a quiescent to an activated state promoting inflammatory responses, chemokine and adhesion molecule expression, and subsequent interaction with platelets and leukocytes.48 o Smooth muscle dysfunction, as showed by a reduced coronary blood flow response to endothelium-independent vasodilators (e.g. adenosine, dipyridamole, papaverine) in patients with microvascular angina.49 o Microvascular spasm and sympathetic dysfunction, mediated via both α1 and α2adrenoceptors in epicardial conduit arteries and microvessels50,51, e.g. after coronary revascularization.  Extravascular factors o Extramural compression (aortic stenosis, hypertrophic cardiomyopathy, arterial hypertension) o Reduction in diastolic perfusion time (aortic stenosis). However, these mechanisms could not be critically tested, establishing clear and accepted pathophysiological concepts. In fact, there is no adequate information on the relevance of individual mechanisms in general, let alone in individual patients. As such, the stratification of individual patients and the rational development of targeted strategies is underdeveloped, as recently acknowledged in the 2015 William Harvey Lecture on Basic Science at the European Society of Cardiology Congress in London46, dedicated to coronary microvascular dysfunction. 3.2. Coronary microcirculation dysfunction in STEMI Despite the fact that primary PCI is highly successful in restoring epicardial coronary blood flow, when performed in a timely fashion, reperfusion at the myocardial level is not accomplished in a significant proportion of patients (from 20 to 60%, depending on the technique used for its evaluation5). The negative prognostic implications (both on the risk of left ventricle remodelling and on the risk of hard endpoints, including death) associated with coronary microvascular damage have been repeatedly confirmed, whatever non-invasive or invasive indicator of microvascular dysfunction/obstruction is used (Table 2). Introduction 48 in humans in vivo, it would still be an incomplete evaluation. Therefore, microcirculatory function is indirectly assessed using several invasive and noninvasive techniques that enable the measurement of parameters that (under normal circumstances) are strongly dependent on the functional integrity of the coronary microcirculation. Consistent with its primary hemodynamic function, functional techniques for the assessment of the coronary microvasculature rely on the measurement of coronary blood flow, which changes mainly as a result of alterations in vascular tone. Although there are several imaging techniques that allow estimating coronary blood flow, at present, the most definite evaluation of the coronary microcirculation remains invasive. In the following pages, the most used invasive and non-invasive techniques for the evaluation of microvascular coronary circulation in patients with acute ST elevation myocardial infarction are described in short. The index of microcirculatory resistance, used in this study, will be detailed in Section 5. 4.1. Non-invasive methods 4.1.1. ECG ST resolution ECG ST-segment resolution is a simple, cheap and validated tool to evaluate acutely microvascular perfusion134 and has been used for several years in fibrinolysis and primary angioplasty trials.3,4 After primary PCI, incomplete ST resolution has been related to coronary microvascular obstruction and worse clinical outcome.135 Different methods have focused on the assessment of multiple leads or single leads showing maximum ST elevation at baseline and a consensus is still lacking about which leads to analyse, the optimal timing of electrocardiogram analysis, and whether standard ECG or continuous ECG monitoring is preferable. Assessment of single lead ST resolution showing maximum ST elevation at baseline seems to be as accurate as the sum of ST resolution measurements.136,137 Additionally, residual ST-segment elevation was found to be an independent marker of coronary microvascular obstruction138 and a predictor of events after primary PCI.139 Introduction 49 4.1.2. Imaging techniques Positron emission tomography (PET) is a well-established non-invasive technique for the assessment of coronary blood flow140,141, as it allows the determination of absolute regional myocardial blood flow at rest and in response to various stimuli. Myocardial blood flow measurement using PET is achieved by continuous monitoring of the radioactivity emitted by an intravenously administered tracer, in the circulation and the myocardium. The kinetics of radiotracer uptake in the myocardium are derived from time-activity curves in the left ventricular cavity and the myocardium; fitting these time-activity curves with an operational equation provides accurate estimates of myocardial blood flow. Importantly, however, it may lack sensitivity and specificity for the diagnosis of coronary vasomotor dysfunction and, in general, is unable to differentiate between epicardial and microvascular abnormalities.142 Myocardial contrast echocardiography (MCE) uses ultrasound to visualize contrast microbubbles with a rheology similar to that of red blood cells that freely flow within patent microcirculation while lack of intra-myocardial contrast opacification is due to microvascular obstruction that predicts functional recovery after STEMI.54 It has proven to be a useful tool for identifying patients with the no-reflow phenomenon after interventional or thrombolytic treatment for STEMI.54,143–145 However, it has several limitations: moderate spatial resolution, operator dependency, and incomplete left ventricular coverage with suboptimal visualization of the lateral wall, and semi-quantitative assessment of coronary microvascular obstruction. Contrast enhanced cardiac magnetic resonance (ceCMR) allows multislice imaging with high tissue contrast and high spatial resolution, enabling accurate quantification of coronary microvascular dysfunction and obstruction, and infarct size. Coronary microvascular dysfunction and obstruction appear as an absence of gadolinium enhancement during first pass and lack of gadolinium enhancement within a necrotic region (late gadolinium hyperenhancement). Coronary microvascular obstruction evaluated by ceCMR correlates with MCE, and other angiographic and invasive indexes146 and is an independent predictor of adverse clinical outcome, alone or adjusted by other factors, such as infarct size and left ventricular ejection fraction.62,147–150 Other imaging modalities, including CT-derived coronary flow reserve and hybrid positron emission tomography-computed tomography are still mainly investigational and have not been widely used in STEMI patients. Finally, periungueal capillaroscopy is a simple and reliable non-invasive technique allowing evaluation of cutaneous microcirculation. However, it has been used mainly in patients with Raynaud’s phenomenon and in patients with connective tissue diseases, and there is no experience with this technique in patients with coronary artery disease. 151 Introduction 50 4.2. Invasive methods 4.2.1. Angiographic indexes Invasive TIMI flow, TIMI frame count and TIMI myocardial perfusion grade have all been proposed in the evaluation of microcirculation after acute ST elevation myocardial infarction.  The thrombolysis in myocardial infarction (TIMI) score grading system describes the rate of blood flow in the epicardial vessels, ranging between no flow at all (Grade 0) to a normal flow rate (Grade 3)152,153 (for more details, see Population and Methods, Section 3.4.1 - TIMI flow grade, page 89). TIMI flow <3 is a marker of both coronary microvascular obstruction and of larger infarct size and has been shown to affect prognosis both at short and long-term follow-up.58 However, the value of this index is limited, since coronary microvascular obstruction may occur in nearly 35% of patients with TIMI 3 flow.146  The corrected TIMI frame count (cTFC) index corresponds to the number of frames required for contrast medium to reach a standardized distal landmark (see Population and Methods, Section 3.4.2 - Corrected TIMI frame count, page 89). It further stratifies the prognosis of patients with TIMI flow 3 and correlates with invasive assessment of coronary flow reserve.154  The TIMI myocardial perfusion grade (TMPG) and the myocardial blush grade (MBG), assess the intensity of the radio-opacity of myocardial tissue after an epicardial coronary injection of contrast medium (MBG), as well as its wash-out rate (TMPG). An intense myocardial blush and fast wash-out of contrast medium indicate optimal microvascular reperfusion.60,155 Both are scored on a scale of 0–3, the latter indicating optimal perfusion (see Population and Methods, Section 3.4.3 - TIMI myocardial perfusion grade (TMPG), page 90). A MPG grade 0–1 can be observed in up to 50% of patients with TIMI flow grade 3.57 Using both TIMI flow and TIMI myocardial perfusion grade156, coronary microvascular obstruction can be defined as a TIMI flow grade <3 or a TIMI flow grade 3 with a TMPG/MPG 0 to 1. 4.2.2. Doppler wire-derived coronary blood flow reserve Coronary flow reserve (CFR) represents the extent to which the coronary circulation can increase myocardial blood flow in response to exercise or a hyperaemic stimulus. It is calculated by dividing hyperaemic flow by resting flow. Accordingly, the calculation of CFR Introduction 51 assumes that maximal vasodilatation is achieved by abolishing coronary vasomotor tone, by intravenous administration of endothelium-independent vasodilators (mostly adenosine). CFR evaluates simultaneously the epicardial and microcirculatory compartments of the coronary tree, although, theoretically, in the absence of epicardial vessel disease CFR reflects microvascular function.157 Invasive evaluation of CFR was first performed with a Doppler guidewire, positioned into the distal part of the coronary artery. With the sensor at the tip of this wire, coronary flow velocity at rest and during hyperaemia can be measured, and the ratio of maximum to baseline coronary flow velocity can be calculated.158 However, this technique has several problems, including frequent guidewire positional changes (motion of the patient, or breathing), which disturbs the Doppler signal and limit the feasibility of the procedure to less than 70% of all arteries.159 Other technical pitfalls that can hamper signal acquisition include obstruction of the guiding catheter, inaccurate calibration, turbulent flow, and signal loss.160 Still, in patients with ST elevation myocardial infarction, CRF evaluated with a Doppler wire was shown as a good prognostic marker for LV function recovery after anterior myocardial infarction treated with primary PCI.52,161 More recently, the prognostic value of microvascular function as determined using Doppler wire for predicting long-term cardiac mortality was assessed in both infarct-related and reference coronary arteries in patients immediately after primary PCI for ST-segment elevation myocardial infarction. During follow-up (median 11 years), a CFR <2.1 in a reference vessel was associated with a fourfold increased risk of longterm cardiac mortality, whereas a target-vessel CFR <1.5 was associated with an increase in short-term (but not long-term) risk of cardiac mortality.63 4.2.3. Pressure wire-derived coronary blood flow reserve In an attempt to overcome the limitations of the Doppler wire, thermodilution-derived CFR was introduced. For this technique, a pressure wire is used to measure temperatures, and CFR is calculated using the principle of thermodilution: by giving short manual injections of 3 cc saline at room temperature into the coronary artery, thermodilution curves are generated and mean transit times at hyperemia and baseline can be calculated. Because coronary flow is inversely proportional to the mean transit time of a bolus of cold saline needed to travel down the coronary artery, CFR can be easily calculated using the ratio of mean transit times. With this technique, successful measurement of CFR can be performed in 95% of patients.159 CRF measurements performed this way correlated well with standard CFR, both in experimental model and in humans.158,162 However, CFR (measured both by Doppler-wire or pressure-wire) varies with age and sex in healthy humans163,164, making it impossible to define a clear cut-off value below which microvascular function could be deemed abnormal. Additionally, since coronary blood flow Introduction 52 under resting conditions is dependent on determinants of myocardial oxygen demand (namely heart rate, contractility and ventricular load) and CFR being the ratio of peak hyperaemic-to-resting flow, it is therefore affected by these same determinants, a fact that can also affect the reproducibility of the ratio.165 CFR it is also significantly influenced by epicardial vessel disease and thus does not distinguish epicardial from microcirculatory disease.158 For all these reasons, CFR alone is not a good tool for evaluating the microcirculation in ST elevation myocardial infarction patients. 4.2.4. Index of microcirculatory resistance To overcome the limitations of CFR and allow for the isolated evaluation of the microcirculatory coronary compartment, a new index, the index of microcirculatory resistance (IMR) was developed.166 The IMR is based on the assumption that microvascular resistance (that provides an independent assessment of microcirculatory function) can be calculated by dividing the distal coronary pressure by absolute coronary flow. According to the methodology previously described for CFR, IMR can be calculated as the distal coronary pressure divided by the inverse of the hyperaemic mean transit time (a correlate to absolute flow), measured simultaneously with the coronary pressure wire. The fundaments, methodology and clinical evidence for this technique will be further developed in Section 5. 4.2.5. Doppler and Pressure combined indexes – hyperaemic microvascular resistance index (hMR) and zero-flow pressure (Pzf) Recently, a single dual sensor wire (Doppler and pressure) was made available (Combo Wire, Volcano Therapeutics), allowing for the simultaneous measurement of phasic distal pressure and flow velocity. With this wire and dedicated software, and using measurements performed at baseline and at hyperaemia induced by adenosine, both hMR (defined as the ratio of average coronary distal pressure and average instantaneous peak velocity during hyperaemia) and Pzf (which is the distal coronary pressure when theoretically the flow in a coronary artery would cease; since it is not possible to measure this directly, as in vivo coronary flow does not cease under normal circumstances, Pzf is extrapolated from pressurevelocity loops167) can be measured. The reproducibility of repeated hyperaemic resistance parameters derived from distal pressure and velocity measurements was confirmed in patients with stable angina.168 In patients with acute STEMI, a few studies were performed, with promising results: In 2003, Shimada and colleagues169 studied Pzf using a Doppler wire and assessing arterial pressure from the guiding catheter in patients undergoing primary PCI for anterior STEMI, and showed Introduction 53 that Pzf correlated with viability evaluated by PET. In 2007, it was confirmed that Pzf values are increased after myocardial infarction.170 More recently hMR, Pzf, and Doppler coronary flow reserve were measured with the combo Doppler-pressure wire in patients after anterior STEMI. All measures correlated with peak creatinine kinase-myocardial and CMR measures of infarct size, and Pzf was found to be higher in those with >75% infarct transmurality.171 In 2015, hMR and Pzf were found to be related to microvascular injury (assessed by ceCMR) and myocardial perfusion (evaluated with PET).172 In the same year, another study compared Pzf, hMR and IMR, and suggested that the former was a better predictor of the extent of myocardial infarction (assessed with ceCMR) than the latter two.173 There is only one study reporting clinical outcomes in STEMI patients treated with primary PCI, and it also confirmed hMR as a strong predictor of a combined endpoint of death and hospital admission for heart failure.174 However, unlike other wire-based techniques, hMR and Pzf have not been validated in animal or human models. Furthermore, there are concerns that hMR measurement may not be accurate in patients with severe epicardial stenosis, since it does not account for collateral flow (and hence may lead to an overestimation of true microvascular resistance in the presence of a severe stenosis). 5. The index of microcirculatory resistance (IMR) 5.1. Definition and evaluation of IMR The IMR is a measurement of the minimum achievable microcirculatory resistance in a target coronary artery territory and thus it provides a quantitative assessment of the microvascular integrity. Unlike CFR, which provides a combined assessment of both the epicardial and the microvascular beds, IMR enables a specific quantitative assessment of the status of the microvascular coronary circulation (Figure 2). Figure 2 – Schematic of physiological assessment (FFR, CFR and IMR) using a coronary pressure and temperature wire Epicardial coronary artery Microcirculatio n Pressure Wire Pd Pa FF R IM R CF R Introduction 54 IMR measurement was made possible by the development of a wire with both pressure and temperature sensors on its tip.158,175 The distal sensor of the wire measures pressure and temperature and the shaft of the wire can be used as a second thermistor. By injecting room-temperature saline down a coronary artery, an indicator-dilution curve is obtained (according to the decrease in temperature) and the mean transit time (𝑻𝒎𝒏) can be determined. 𝑻𝒎𝒏 has been shown to strongly correlate inversely with absolute flow and therefore provides an easily measurable surrogate, both in vitro, in animal models158 and in humans.162 According to Ohm’s law, the resistance (𝑹) in a given circuit is related to the pressure gradient (∆𝑷) and absolute flow (𝑸), according to the formula: 𝑹 = ∆𝑷 𝑸 In the coronary microcirculation, the pressure gradient (∆𝑷) is determined by distal pressure (distally in the epicardial vessel) minus venous pressure; at maximal hyperaemia, however, venous pressure can be assumed to be zero, implying that ∆𝑷 approximates distal pressure (Pd), measured with the pressure-wire. Similarly, at maximal hyperemia, absolute flow (𝑸) has been shown to have a linear relationship with the inverse of flow velocity (𝟏𝑻𝒎𝒏 ⁄). Therefore, the microcirculation resistance can be derived from the simplified formula176: 𝑰𝑴𝑹 = 𝑷𝒅 (𝟏𝑻𝒎𝒏 ⁄ ) Or, to make it even simpler, 𝑰𝑴𝑹 = 𝑷𝒅 ∗ 𝑻𝒎𝒏 Importantly, IMR is derived from the assumption that, at peak hyperaemia, the variability of resting vascular tone and hemodynamic will be eliminated, and the minimum microvascular resistance will be achieved. In the presence of a severe epicardial stenosis, myocardial flow is the composite of both the coronary and collateral flows. Accordingly, distal coronary pressure decreases to a lesser degree, given the contribution of collaterals – this in turn will produce an overestimation of IMR if not corrected for collateral flow. Therefore, to calculate the true IMR (IMRtrue) in this setting, a more complex formula has been developed, which includes measurement of the coronary wedge pressure as a measure of collateral pressure.177 However, in the absence of significant collateral flow, it is reasonable to use IMR as a surrogate for IMRtrue, avoiding the measurement of coronary wedge pressure. IMR evaluation was validated in animals and showed a strong correlation with true microvascular resistance, being independent on epicardial stenosis presence and severity (since both distal Introduction 55 pressure and flow drop in the presence of an epicardial stenosis).166 This was also validated in humans.177,178 Despite the relatively complex theoretical background, IMR is easy to evaluate and a very reproducible measurement.179 A dedicated console, equipped with dedicated software for IMR evaluation makes the procedure simple and easy to learn. The complete description of the procedure is given in detail in Population and Methods, Section 3.5 – Index of microcirculatory resistance (IMR), page 90. It is a safe technique to perform in STEMI patients180 and has been used in several studies (described in the next section). 5.2. Clinical evidence on the use of IMR in STEMI patients The evaluation of coronary microcirculation involvement and its consequences in patients with acute STEMI treated with primary angioplasty is a logical application of IMR. Indeed, several studies were performed in ST elevation acute myocardial infarction patients, immediately after the primary PCI. These studies, summarized in Tables 3-6, showed that IMR is related with:  Left ventricular remodelling (either evaluated by wall motion score index, improvement in left ventricular volumes or improvement in left ventricle ejection fraction) between the acute phase and follow-up, using either echocardiography or contrast enhanced cardiac magnetic resonance (ceCMR),  Infarct size, measured either by the amount of cardiac biomarkers released, by singlephoton emission computed tomography (SPECT) or by ceCMR,  The presence of microvascular obstruction, evaluated by ceCMR, and  Myocardial viability, assessed by positron emission tomography (PET). Additionally, recent evidence suggests that IMR measured immediately after the primary PCI in STEMI patients is a strong and independent predictor of major events, including death (Table 5). Although almost all of these studies were unicenter and small (the majority included less than 5060 patients), their results, coming from different groups of investigators, are remarkably consistent. Therefore, in the last years, IMR has also been used as a tool to evaluate both pharmacological strategies (like IIb/IIIa inhibitors, nicorandil or nitroprusside) and devices (like thrombus aspirators or distal protection devices) in the treatment of STEMI patients (Table 6). Finally, IMR has been widely used in stable patients, for several indications176,181: as an adjuvant of FFR in patients with equivocal clinical presentation and intermediate coronary stenosis182, in patients with probable microvascular angina183, to evaluate drugs184–186 or devices187, to assess the impact on microcirculation of percutaneous revascularization188,189, and in other specific clinical conditions (apical ballooning syndrome190, assessment of steam cell therapy191,192, transplant arteriopathy193 etc.). However, to the date, there are no studies relating IMR to endothelialdependent tests in patients with acute myocardial infarction. Introduction 56 Table 3 – Studies using IMR to evaluate LV remodelling in patients with ST elevation myocardial infarction Author (Year) No. of patients Comparator Outcome Results LV remodelling Fearon (2008)238 28 Echocardiography WMS at 3 months Significant correlation (r = 0.59, p = 0.002) IMR>32 associated with significantly worse WMS IMR was the only independent predictor of WMS Lim (2009)249 40 Echocardiography Change in anterior wall motion score at 6 month Significant correlation (r = -0.464, p = 0.003) Cut-off 33U IMR: AU ROC for predicting LV recovery: 0.89 (CI 95% 0.888-0.894) Sezer (2010)250 35 Echocardiography LV volumes and LVEF at 5 months (IMR measured acutely and at 5 months); significant superior improvements in LV volumes and LVEF in patients that increased more IMR between the 2 evaluations McGeoch (2010)251 47 Contrast enhanced CMR LV volumes and LVEF (2 days and 3 months) IMR was an independent predictor of LVEF at 2 days and 3 months Yoo (2012)252 34 Echocardiography Change in WMSI at 6 months Change in LVEF at 6 months Significant correlation (r = -0.61, P < 0.01) with change in WMSI Significant correlation (r = -0.52, P < 0.01) with change in LVEF Faustino (2016)56 40 Echocardiography Change in WMSI, LVEF and E/e’ between acute and 3 months All parameters significantly improved more in patients with lower IMR Lower IMR was associated with lower acute global longitudinal strain values Carrick (2016)290 259 Contrast enhanced CMR Change in LV end-diastolic volume and in LVEF at 6 months IMR was an independent predictor of LV end-diastolic and LVEF improvement between ceCMR performed at 2 days and 6 month, but CFR was not Park (2016)257 89 Echocardiography WMSI at 3 months Improvement in patients with lower IMR and higher FFR groups, but not on patients with both higher IMR and lower CFR Palmer (2016)258 31 Echocardiography Change in WMSI and LVEF at 3 months (treatment = thrombolysis) IMR correlated with LVEF (R = 0.652, p=0.005) and WMSI (R = 0.452, p=0.011) Only patients with lower IMR showed LV recovery at 3 months Introduction 57 Table 4 – Studies using IMR to evaluate the size of the infarction, microvascular obstruction and myocardial viability in patients with ST elevation myocardial infarction Author (Year) No. of patients Comparator Outcome Results Extension of the infarction Fearon (2008)238 32 Cardiac biomarkers CK peak values Significant correlation with peak CK (R = 0.61, p = 0.0005) IMR>32 associated with significantly higher peak CK values IMR was the strongest independent predictor of peak CK Sezer (2010)250 35 Tec-99m sestamibi SPECT Infarct size (IS) (at 2 days and 5 months) IMR was a predictor of IS at 5 months (β = 0.280, P = 0.013) IMR was a predictor of change in IS between 1st and 2nd SPECT (r=0.55, P=0.001) In patients with an improvement in IMR >33%, IS decreased significantly McGeoch (2010)251 47 Contrast enhanced CMR Infarct volume (2 days and 3 months) IMR was an independent predictor of infarct volume at 2 days and 3 months Payne (2012)253 96 Contrast enhanced CMR Infarct volume and salvage index (2 days and 3 months) IMR was an independent predictor of infarct volume and salvage index (the difference between infarct size and area at risk) at 2 days, but not at 3 month Carrick (2016)290 281 Contrast enhanced CMR Infarct size, % LV mass Higher IMR values were associated with larger infarct size Microvascular obstruction McGeoch (2010)251 53 Contrast enhanced CMR Microvascular obstruction (MVO) IMR significantly higher in patients with MVO Yoo (2012)252 34 Contrast enhanced CMR Microvascular obstruction (MVO) MVO correlated with IMR (r = 0.754; P < 0.001) Payne (2012)253 96 Contrast enhanced CMR Microvascular obstruction (MVO) IMR was an independent predictor of MVO (at 2 days) Fukunaga (2014)254 88 Contrast enhanced CMR Microvascular obstruction (MVO) Patients with bimodal shape IMR curve had significantly higher MVO Ahn (2016)248 40 Contrast enhanced CMR Microvascular obstruction (MVO) IMR and CFR both good predictors of MVO, but the combination of both was even better for predicting MVO (AUC 0.941) Carrick (2016)290 219 Contrast enhanced CMR Microvascular obstruction (MVO) IMR was an independent predictor of MVO (at 2 days) Myocardial Viability Lim (2009)249 38 PET (18Ffluorodeoxyglucose) Regional myocardial FDG uptake at 6 months Significant correlation (R = -0.738, p < 0.001) Introduction 64 within the tonometer.213 The pulse-waveform shape obtained provides information about arterial compliance (including the augmentation index, a ratio between the pulse pressure at the second systolic peak and the pulse pressure at the first systolic peak, which is commonly used as a measure of arterial stiffness) and about endothelial function (by measuring the changes in the peripheral pressure waveform in response to β-2 adrenergic stimulation).214,215 However, this technique has rarely been used to investigate endothelial dysfunction in clinical settings. 6.2.4. Peripheral (pulse) arterial tonometry In addition to applanation tonometry, pulse wave amplitude of the peripheral microvasculature can be assessed by measuring changes in digital pulse volume using a finger photopletysmograph (pulse contour analysis).216 Reactive hyperemia peripheral artery tonometry is a recent development of this technique, which uses plethysmography to record digital volume changes accompanying pulse waves.217 Its fundaments and clinical evidence will be further detailed in the next section. 6.2.5. Laser Doppler flowmetry of the skin Laser Doppler flowmetry (LDF) is a technique that enables the monitoring of skin microvascular blood flow.218 The assumption is that the response observed in the cutaneous circulation is a window towards the responses that would be observed in other vascular beds.219 During LDF, the original beam of coherent light changes in contact with moving tissues (red blood cells) and a photodiode measures the emerged beam. The fraction of shifted light depends on the concentration of moving red blood cells, whereas the magnitude of the frequency broadening depends on their average velocity.220 LDF has been used to evaluate endothelial function of the skin microvasculature using postocclusive hyperemia, local thermal hyperemia, and acetylcholine iontophoresis. Despite being noninvasive and, therefore, attractive for routine clinical and research use, these techniques have certain limitations. Firstly, because the skin is a critical thermoregulatory organ, there are extreme variations in basal blood flux, which, in turn, dictate the need to use maximal vasodilatation (by either local warming of the skin or local sodium nitroprusside infusion) to normalize submaximal flux values. Secondly, poor interassay and intra-assay reproducibility and lack of standardization (e.g., site of the skin measurement) limit withinpatient and across-studies comparisons. Thirdly, and most importantly, recent insights into the mechanisms of the postocclusive hyperemia and acetylcholine-mediated dilatation indicate that these phenomena are not primarily NO mediated, suggesting that they might represent a summation of complex, microvascular responses involving sensory nerves and Introduction 65 metabolic and endothelial vasodilators (independent from NO). Therefore, rather than representing specific markers of endothelial function, these tests provide a more global form of assessing microvascular function.41 6.2.6. Biochemical biomarkers Markers of coagulation/thrombosis The plasma levels of several procoagulant mediators have been shown to increase with endothelial damage, suggesting that they could represent reliable markers of endothelial dysfunction. Furthermore, a change in the production of these molecules by the endothelium could directly contribute to atherotrombotic disease. These markers include, among others, von Willebrand factor, tissue plasminogen activator (t-PA) and PAI-1. However, prospective epidemiological studies aiming to evaluate the association between plasma levels of different hemostatic markers and the risk for cardiovascular disease are relatively sparse and often have inconclusive results.41 Markers of inflammation Strong evidence suggests that atherosclerotic risk factors are often associated with systemic inflammation, which is a key player in the development and progression of atherosclerosis.106 C-reactive protein, in particular, has emerged as a potential marker for cardiovascular risk. It can be measured with several standardized, validated, and inexpensive high-sensitivity assays and is the only biomarker ready for clinical use, adding predictive value above the currently established risk factors, both in stable and unstable patients.41 However, it is affected by several mechanisms and factors, which limits its use as a marker of endothelial function. Other markers of endothelial dysfunction Asymmetrical dimethylarginine (ADMA) and oxidized low-density lipoprotein (oxLDL) are two other markers which have been investigated as indicators of endothelial function. However, despite initial promising results, suggesting an association with cardiovascular disease and risk, both had more recent conflicting results and there are doubts if they represent reliable substitutes for direct endothelial function measurement.211 Endothelial microparticles and bone-marrow derived endothelial progenitor cells have also been widely investigated as markers of endothelial dysfunction, with promising results, but they remain mostly investigational tests.211 Finally, baseline plasma levels of several cellular adhesion molecules (CAMs), including intercellular adhesion molecule-1 and E-selectin, have been shown to be associated with increased cardiovascular risk in generally healthy population. In patients with coronary artery Introduction 66 disease, elevated circulating vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 values were also predictors of adverse outcome. However, the results from other large studies indicated that, after adjustments for other cardiovascular risk factors, the association of CAMs with coronary heart disease is not statistically significant.41 7. Peripheral (pulse) arterial tonometry (EndoPAT) 7.1. Principle and methodology Reactive hyperemia peripheral artery tonometry uses plethysmography to record digital volume changes accompanying pulse waves.217 On the basis of this principle, a finger pneumatic plethysmographic cuff, providing a ‘‘beat to beat’’ blood flow volume assessment by recording finger (peripheral) arterial pulsatile (PAT) volume changes, was developed by Itamar Medical, and made commercially available under the name EndoPAT 2000® (Figure 9, page 95). It consists of a fingertip plethysmograph capable of sensing volume changes in the digit with each arterial pulsation. The fingertip probe has a rigid external casing containing inflatable chambers and the uniformly applied pressure field across the finger prevents venous pooling and partially unloads arterial wall tension (Figure 10, Page 96). Volume changes in the fingertip are recorded digitally as pulse amplitude that can be tracked over time. Endothelial function testing with PAT is based on the same physiological mechanisms as the FMD technique, inducing transient ischemia in the upper limb as a stimulus for reactive vasodilatation:  A pressure cuff is placed around the arm and inflated above systolic pressure after a baseline recording  The cuff is then deflated after 5 minutes to induce reactive hyperemia in one arm (a main advantage of the system is that the contralateral arm serves as an internal control).  The ratio between hyperaemic and baseline pulse volume analysis (PAT ratio) is normalized for the same ratio in the contralateral arm, thus obtaining the reactive hyperemia index (RHI), or its natural logarithm (L_RHI).221 This index is a validated marker of endothelial function – a value below 1.67 (or ≤0.51 for L_RHI) is considered suggestive of endothelial dysfunction. The reproducibility of the procedure was also clearly established.222–224 The technical details of the EndoPAT procedure are further detailed in Section 3.9 of the Methods Chapter (Page 95). Unlike flow-mediated dilation evaluation, the PAT technique is operator-independent, and the pulse amplitude recordings are digitized and analysed by an automated, proprietary algorithm. Introduction 67 7.2. Clinical evidence on the use of EndoPAT in coronary artery disease The role of EndoPAT in the evaluation of patients with coronary artery disease is supported by the study performed by Bonetti and colleagues, in which they showed that a lower PAT hyperaemic response is significantly correlated with the presence of coronary endothelial dysfunction measured by acetylcholine response207 in patients undergoing coronary angiography. This study confirms previous evidence endorsing the concept of a generalized nature of endothelial dysfunction206,208 and supports the use of endothelial dysfunction evaluated by EndoPAT as a surrogate for coronary endothelial dysfunction. Additionally, EndoPAT has been evaluated in different stages of coronary artery disease:  In patients without known/suspected coronary artery disease, PAT hyperaemic ratio is progressively lower with increasing burden of cardiovascular risk factors225–227 and several studies confirmed that endothelial dysfunction evaluated with EndoPAT is related both to the risk of developing coronary disease and the risk of cardiac events. In fact, as described in Section 3.3 above and also confirmed with this technique, the presence of endothelial dysfunction appears to have an incremental prognostic value, after control for these classical risk factors, which may be explained by the fact that it reflects the overall burden of risk and significantly predicts the risk of cardiovascular events. The main prognostic studies performed with EndoPAT to evaluate the risk of events in patients without coronary artery disease are presented in Table 7.  In patients with known stable coronary artery disease, endothelial dysfunction evaluated by EndoPAT was related with the risk and severity of the disease and to the characteristics of coronary plaques (Table 8). EndoPAT was also used to evaluate the effect of enhanced external counter pulsation in patients with stable angina and refractory complaints. Additionally, three long-term studies performed with EndoPAT confirmed its additional and independent value in predicting events in patients with coronary artery disease (Table 9).  Finally, in patients with acute coronary syndromes and ST elevation myocardial infarction, evidence with EndoPAT includes studies predicting the risk of in-stent restenosis, initial patency of the culprit artery, angiographic severity of the coronary artery disease and the extension of infarction in STEMI patients (Table 10). These two last studies were performed in our centre, and they will be further described below, since they were essential for the design of this thesis protocol. Introduction 68 Table 7 – Prospective studies investigating the predictive role of peripheral arterial tonometry in patients with suspected coronary artery disease or other clinical settings Author, Year No. of patients Clinical Context FUP duration Clinical endpoints Results Rubinstein (2010)129 270 Suspected CAD 5.8 years CV death, nonfatal MI, coronary revascularization or CV hospitalization Low L_RHI(<0.4) associated with higher rate of events (48% vs. 28%, p=0.03) L_HRI was an independent predictor of adverse events (HR 1.79 CI95 1.162.76, P=0.008) Akiyama (2012)275 321 Heart failure with preserved ejection fraction 20 months CV death, nonfatal MI, unstable angina pectoris, nonfatal ischemic stroke, coronary revascularization, hospitalization for HF RHI (HR 0.80, 95% CI 0.67-0.94, P=0.007, for an increase in RHI of 0.1) was an independent predictor of events. RHI improved the prediction of events (C-statistics from 0.671 to 0.712) when added to 5 other risk factors (age, diabetes, NYHA classification, heart failure hospitalization history, and LVEF) Matsue (2013)276 159 Heart failure with preserved ejection fraction 300 days HF-related death, hospitalization for HF L_RHI (HR 0.56, 95% CI: 0.39-0.80 for an increase of 0.1) was an independent predictor of HF-related events. AUC of ROC curve for L_RHI was 0.73 (95% CI: 0.62–0.83) L_RHI 0.49 was suggested as the optimal cut-off value for prediction of adverse events in this population. Matsuzawa (2013)277 528 Suspected CAD 2.8 years CV death, nonfatal MI, unstable angina, nonfatal ischemic stroke, coronary revascularization, hospitalization for HF, nonfatal aortic disease, and peripheral arterial disease RHI (HR 0.761 95%CI 0.673-0.859; P<0.0001, for an increase of 0.1) was an independent predictor of events RHI improved the prediction of events (C-statistics from 0.728 to 0.766) when added to 3 other risk factors (FRS, SYNTAXsc and BNP) Hirata (2014)278 383 Chronic kidney disease 30 months CV death, non-fatal MI, unstable angina, nonfatal ischemic stroke, coronary revascularization, hospitalization for HF Low L_RHI (HR 2.70, 95% CI: 1.62-4.51, p<0.001) was an independent predictor of events RHI improved the prediction of events (C-statistics from 0.49 to 0.62) when added to FRS. Introduction 69 Table 8 – Prospective studies investigating peripheral arterial tonometry in patients with stable coronary artery disease Author, Year No. of patients Clinical Context Objectives Results Bonetti (2003)222 23 Stable CAD, refractory angina Effect of enhanced external counterpulsation (EECP) RHI significantly increased only in patients that clinically improved (less symptoms) with EECP Bonetti (2010)323 140 Stable CAD Prediction of obstructive and non-obstructive CAD RHI significantly attenuated in both obstructive (median 1.57, IQR 1.42-1.76) and nonobstructive CAD (median 1.58, IQR 1.41-1.78), as compared to non-IHD (2.15, IQR 1.852.48, p < 0.001). RHI (HR 0.51 95%CI 0.38-0.68; P<0.001) was an independent predictor of CAD AUC ROC curve for prediction of CAD: 0.86, p<0.001 Heffernan (2010)324 42 Stable CAD Identification of patients with high risk CAD (elevated hsCRP and Lp-PLA2) RHI significantly lower in the high-risk group (1.3 ± 0.04) compared to the moderate-risk (1.6± 0.07, P < 0.05) and low-risk (2.0± 0.1, P < 0.05). RHI was a significant independent predictor of high risk CAD Schonenberger (2012)267 362 Stable CAD IVUS-assessed coronary plaque structure. Patients with RHI<1.67 had greater plaque burden (41% vs 39%, p = 0.047) RHI significantly associated with fibrous and fibrofatty plaques and with necrotic core and dense calcium volumes Introduction 70 Table 9 – Prospective studies investigating the predictive role of peripheral arterial tonometry in patients with established coronary artery disease Author, Year No. of patients Clinical Context FUP duration Clinical endpoints Results Suessenbacher (2014)325 96 Stable CAD 44 months Revascularization, acute coronary syndrome, ischemic stroke, cardiovascular death, repeat coronary angiography due to chest pain Risk similar in patients below median RHI. However, the combined endpoint occurred earlier in the patients with an RHI within the 1st tertile (11.3±11.0 vs. 27.5±18.6 months for patients in the 2nd/3rd tertile, p=0.03) Matsue (2014)326 213 Stable CAD, under statin treatment, LDL-C< 100 mg/dl 2.7 years Angina pectoris requiring coronary revascularization, recurrent angina pectoris with proven myocardial ischemia, non-fatal MI, and death from CAD L_RHI (HR 0.79, 95% CI: 0.66-0.95, p=0.012) was an independent predictor of events RHI improved the prediction of events (C-statistics from 0.60 to 0.77) when added to risk factors and estimated glomerular filtration rate. Tabata (2015)327 435 PCI patients with chronic kidney disease (CKD) 2.5 years Cardiovascular death, nonfatal myocardial infarction, ischemic stroke, hospitalization due to unstable angina pectoris, and coronary revascularization L_RHI lower in patients with events in the non-CKD group (0.46 ± 0.18 versus 0.60 ± 0.25; P = 0.002) L_RHI (HR: 0.096; 95 % CI 0.02–0.47; P = 0.004) was an independent predictor of cardiovascular events in the non-CKD group Introduction 71 Table 10 – Prospective studies investigating peripheral arterial tonometry which include patients with acute coronary syndromes or ST elevation myocardial infarction Author, Year No. of patients Clinical Context Objectives Results Baptista (2013)19 231 Stable angina and ACS patients Severity of CAD RHI was progressively lower as CAD severity increased (from no disease to 3 vessel disease: 1.98±0.46 -> 1.86±0.46 in 1 vessel -> 1.85±0.43 in 2V -> 1.60±0.39 in 3V, p=0.003) RHI (HR=0.16 for each unit of RHI reduction , 95%CI 0.04-0.68, p=0.013) was independent predictor of severity of CAD Yamamoto (2014)268 86 STEMI patients Risk of in-stent restenosis (ISR) RHI at 8 months significantly lower in the patients with ISR (1.75 vs. 2.12; p=0.03). RHI (OR: 4.23, 95% CI 1.25-14.28, p=0.02) was a significant risk factor for ISR Kandhai-Ragunath (2014)269 71 STEMI patients Initial patency of culprit artery RHI significantly higher if culprit vessel patent before P-PCI (2.08±0.34 vs. 1.75±0.35; p<0.007). RHI (OR 7.1, 95% CI 2.1-23.6) was an independent predictor of culprit vessel patency Baptista (2016)18 58 STEMI patients Extension of infarction (peak TnI) Patients with an RHI<1.67 had significant larger infarcts (TnI 73.5 [IQR 114.42] vs. 33.2 [IQR 65.2 ng/mL]; p =0.028) RHI was an independent predictor of peak TnI values Komura (2016)328 249 Stable angina and ACS patients Risk of in-stent restenosis RHI_L at follow-up lower in patients with ISR (0.52±0.23 versus 0.65±0.27, P<0.01) RHI_L (OR 0.13; 95% CI: 0.04-0.48; P=0.002) was a significant independent predictor of ISR Introduction 72 7.2.1. Severity of coronary artery disease and endothelial dysfunction evaluated by peripheral arterial tonometry19 With the purpose of evaluating the prevalence of endothelial dysfunction by severity of coronary artery disease, 231 patients referred for diagnostic angiography and 39 control subjects, were evaluated by peripheral arterial tonometry (EndoPAT) in our department. The severity of coronary artery disease was defined as the number of vessels with disease (lesions >70%). Of the 231 patients, 92 (39.8%) had no relevant disease (“normal” coronaries), 78 (33.8%) had 1 vessel disease, 37 (16.0%) had 2 vessel disease and 24 (10.4%) had 3 vessel disease. Mean RHI values in the control group were 2.10±0.63. In catheterised patients, RHI was progressively lower as CAD severity increased: 1.98±0.46, 1.86±0.46, 1.85±0.43 and 1.60±0.39, respectively in patients with “normal” coronaries, 1, 2 and 3 vessels disease (p=0,003). Since patients with 1 and 2 vessels had similar RHI results, they were merged in one group for the analysis performed (Figure 3, Table 11). Table 11 – Patient characteristics according to the severity of coronary artery disease a presented as N (%); b Presented as mean±standard deviation: c p-value for the comparison normal coronaries vs. 1-2 vessel disease vs. 3 vessel disease: One-Way ANOVA for continuous variables and Chi-Square for categorical variables. d p=0.002 (OneWay ANOVA; Controls vs. normal coronaries vs. CAD) Three vessel disease was more prevalent in male patients (in the unadjusted analysis). Patients with multivessel disease also tended to be older and more often diabetics. The other characteristics were similar between groups. Controls “Normal“ coronaries 1-2 vessel disease 3 vessel disease P value c Patients a 39 92 (39.8%) 115 (49.8%) 24 (10.4%) - Reactive hyperemia index b 2.08 ± 0.63 1.98 ± 0.46 d 1.85 ± 0.45 1.60 ± 0.39 0.001 Physical characteristics Age (years) b 39.5± 12.8 59.5 ± 13.7 † 60.3 ± 13.7 64.2 ± 13.1 ns Male gender a 17 (40.5%) 44 (47.8%) †† 87 (75.7%) 20 (83.3%) <0.001 Body mass index b 25.4 ± 5.6 27.2 ± 4.1 † 27.2 ± 4.4 28.3 ± 3.9 ns Heart rate b 66.2 ± 9.2 65.0 ± 11.7 † 65.6 ± 11.2 69.1 ± 14.4 ns Systolic blood pressure b 114.1 ± 14.4 121.7 ± 17.3 † 116.7 ± 20.8 118.4 ± 22.6 ns Risk Factors a Diabetes 0 (0.0%) 13 (14.2%) 28 (24.3%) 8 (33.3%) ns Hypertension 6 (14.3%) 66 (72.5%) 77 (67.0%) 18 (75.0%) ns Dyslipidaemia 12 (28.6%) 48 (52.2%) 54 (47.0%) 14 (58.3%) ns Smoking 9 (21.4%) 20 (22.2%) ‡ 47 (40.9%) 9 (21.4%) 0.018 Introduction 73 Figure 3 – Reactive hyperemia index according to the severity of the coronary artery disease (number of major vessels with lesions >70%) In the univariate analysis, comparing 3-vessel disease with 1 or 2-vessel disease, RHI and nonST acute coronary syndrome as indication for the angiography (vs. ST elevation myocardial infarction) were the strongest predictors of the severity of disease (Table 12). However, when this analyses was adjusted (for age, gender, previous diabetes mellitus or dyslipidaemia, BMI and waist circumference), only RHI was independently associated with the diagnosis of 3 vessel disease (Table 13). Table 12 – Univariate analysis of predictors of 3 vessel disease in the population with documented coronary artery disease nST-ACS: non-ST acute coronary syndrome; STEMI: ST elevation myocardial infarction Univariate analysis 1-2 Vessels 3 Vessels RR 95% CI P value Reactive hyperemia index 1,85±0,45 1,60±0,39 0,014 nST-ACS (vs STEMI) 44,3% 70,8% 2,54 1,12-5,73 0,016 Gender (male) 75,7% 83,3% 1,61 0,51-5,11 0,301 Dyslipidaemia 47,0% 58,3% 1,58 0,65-3,85 0,215 Diabetes mellitus 24,3% 33,3% 1,55 0,60-4,02 0,251 Hypertension 67,0% 75,0% 1,48 0,54-4,03 0,304 Age (per year) 60,3±13,6 64,2±13,1 0,193 Waist circumference (cm) 98,5±13,0 104,6±16,3 0,051 Study Objectives 81 Study Hypothesis The study´s main hypothesis is: In patients with acute ST elevation myocardial infarction treated with primary PCI, endothelial dysfunction (evaluated by peripheral arterial tonometry) is related to the extent of microvascular damage and, consequently, to the extent of myocardial necrosis. Based on this hypothesis, the study primary aim is to evaluate the relation between endothelial dysfunction (evaluated with EndoPAT) and coronary microvascular dysfunction (evaluated by IMR immediately after primary revascularization). Secondary aims include both confirming IMR and evaluating endothelial dysfunction (as measured by peripheral arterial tonometry) as predictors of microvascular dysfunction and extension of the myocardial infarction. Study Outcome Measures The primary outcome measure is the IMR value. Since there are no clearly defined values for abnormal endothelial function in coronary artery disease patients, two pre-specified evaluations of this primary outcome measure were defined: 1. IMR value in patients with endothelial dysfunction according to the prevailing threshold defined for EndoPAT (RHI <1.67 versus ≥1.67) 2. IMR value in patients according to the tertile of RHI Secondary outcomes of the study are: 1. The relation between RHI values and a) The extent of myocardial necrosis, evaluated by troponin release (in the first 48 hours), echocardiographic parameters (both measured acutely and at 3 months) and ceCMR; b) The extent of microvascular reperfusion indicators, including angiographic indicators (cTFC and TMPG), ECG (ST resolution), and ceCMR (microvascular obstruction). 2. The relation between IMR values and a) The extent of myocardial necrosis, evaluated by troponin release (in the first 48 hours), echocardiographic parameters (both measured acutely and at 3 months) and ceCMR; b) The extent of microvascular reperfusion indicators, including angiographic indicators (cTFC and TMPG), ECG (ST resolution), and ceCMR (microvascular obstruction) 83 POPULATION AND METHODS Methods 84 1. Type and Location of the Study Observational, prospective, single centre, cohort study, performed in the Cardiology Department of Hospital Prof. Doutor Fernando da Fonseca (Amadora, Portugal). 2. Population All patients admitted to Hospital Prof. Doutor Fernando da Fonseca with a first acute ST elevation myocardial infarction, treated with primary angioplasty, were considered for inclusion in the study, according to the inclusion and exclusion criteria defined below. 2.1. Inclusion Criteria  Age > 18 years.  First ST elevation acute myocardial infarction, defined by: - Chest pain lasting at least 20min and /or - ST segment elevation of at least 1 mm in two or more contiguous leads  Pain to balloon time < 6 hours (or between 6 and 12 hours if clearly with ongoing pain)  Primary PCI performed with success on the culprit lesion (with no significant residual stenosis, independently of the final TIMI flow) in a native coronary artery.  Informed consent obtained 2.2. Exclusion Criteria  Patients presenting with left bundle-branch block and patients with implanted pacemaker – excluded since ST resolution could not be evaluated.  History of previous of myocardial infarction (either ST elevation or non ST elevation MI) – excluded in order to reduce the bias in the ceCMR and in the microvascular tests.  Patients with clear retrograde circulation to the infarct related artery (Rentrop ≥2).  Killip class IV (cardiogenic shock) on presentation or during the primary PCI procedure  Patients with known myocardial diseases (such as hypertrophic cardiomyopathy or restrictive cardiomyopathies) and patients with severe left ventricular hypertrophy (wall thickness > 15 mm) – were excluded in order to reduce the bias in the ceCMR and in the microvascular tests.  Previous coronary artery bypass surgery.  Percutaneous revascularization in the last 3 months. Methods 85  Long QT syndrome, 2nd or 3rd degree heart block and sick sinus syndrome (due to the risk of severe bradycardia/arrhythmias induced by adenosine)  Patients in atrial fibrillation (if ceCMR was to be performed) or with other arrhythmias considered by the investigator as serious enough to contra-indicate the use of adenosine immediately after the primary angioplasty.  Severe asthma or chronic pulmonary obstructive disease (due to the risk of bronchospasm induced by adenosine)  Previous severe reaction to adenosine or any other contraindication to adenosine, including systolic blood pressure < 90 mmHg or bradycardia deemed to be significant by the operator.  Presence of any serious non-cardiac disease associated with a life expectancy of less than 12 months  Inclusion in other trials/studies 2.3. Sample Since there were no clearly defined “normal” values both for RHI and IMR in patients with acute myocardial infarction, sample calculation was a difficult task to accomplish. Our previous work with EndoPAT in patients with ST elevation myocardial infarction18, showed that RHI values lower than 1.67 are associated with larger infarcts (measured by peak troponin I release). Values lower than 1.67 were present in approximately 30% of patients (unpublished data), which translates into an “exposed/non-exposed” ratio of 2.3. In patients with RHI lower than 1.67, approximately 70% had larger infarcts (peak troponin I > 50 ng/dL). Using the OpenEpi sample size calculator (Kelsey method) 228, for a two-sided significance level (α) of 0,05, a power (1-β, % chance of detecting) of 80 and assuming a proportion with disease (larger infarcts) in patients with RHI > 1.67 (“non-exposed”) of 30%, the sample size would be 58 patients (18 “exposed” + 40 “non-exposed”). 3. METHODS 3.1. Clinical information Data concerning patient physical characteristics (age, gender, weight, height, body mass index, waist circumference), major risk factors (high blood pressure, diabetes, dyslipidaemia, smoking habits) and previous coronary/non-coronary atherosclerotic history (previous angina, previous percutaneous revascularization were collected in all patients included. Data on previous drug treatment was also registered. Methods 86 3.2. Primary Percutaneous Coronary Intervention Primary percutaneous coronary angioplasty was performed according to the recommended standards. The time of symptoms onset, time of hospital admission and time of first coronary intervention (thrombus aspiration, balloon dilation or stenting, whichever was first) was recorded, in order to allow the determination of ischemic times:  “pain-to-hospital”: time between the onset of symptoms and the first medical contact;  “hospital-to-balloon”: time between the first medical contact and the first coronary intervention; and,  “pain-to-balloon”: time between symptom onset and the first coronary intervention. The treatment strategy (thrombus aspiration, direct stenting versus balloon pre-dilatation, balloon post-dilatation, type and number of stents, dose of non-fractionated heparin, use of GP IIb/IIIa inhibitors, dose and timing of clopidogrel, etc.) was left to the discretion of the primary operator. All these technical parameters were collected. After the primary PCI, patients were admitted to the Intensive Cardiac Unit and offered standard care for patients with acute STEMI, including recommended pharmacological therapy (aspirin, clopidogrel/ticagrelor, beta-blockers, ACE inhibitors and statins), according to Portuguese229 and European1 guidelines. 3.3. Coronary angiography and area-at-risk scores Angiography was performed according to usual procedures, by one of the 4 senior operators of the Interventional Cardiology Unit of Hospital Prof. Doutor Fernando da Fonseca. All operators were largely experienced in treating patients with acute ST elevation myocardial infarction. The severity of coronary artery disease was evaluated by the Syntax Score.230 Lesion segments were classified (as in the Syntax Score) based on the 16 segment definition proposed by the American Heart Association and modified for the ARTS I and II trials.230 The presence of collateral flow to the culprit artery was evaluated, using the Rentrop collateral flow classification231:  Grade 0: None  Grade 1: Filling of side branches of the artery to be dilated via collateral channels without visualization of the epicardial segment;  Grade 2: Partial filling of the epicardial segment via collateral channels;  Grade 3: Complete filling of the epicardial segment of the artery being dilated via collateral channels Methods 87 Patients with a Rentrop flow 3 or 4 in the infarct related artery were not included, in order to reduce the bias in IMR measurement in this territory. For each lesion, the area-at-risk was calculated, according to the segment involved and using two different scores: the APPROACH score and the BARI score. These 2 scores have been widely validated in clinical practice.232–235 3.3.1. Modified APPROACH Score The Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH) score236 is an angiographic score in which the left ventricle is divided into regions according to the percentage of myocardium supplied by a vessel or its branches. The area-at-risk for a given lesions is calculated taking into account the location of the culprit lesion, dominance and size of the secondary branches. The modified APPROACH score232 simplifies this evaluation, using a table with values defined for each lesion location (Figure 5). Figure 5 – Modified APPROACH score for evaluating the area-at-risk (adapted from Ortiz-Perez et al233) Methods 88 3.3.2. BARI Score The Bypass Angioplasty Revascularization Investigation Myocardial Jeopardy Index (BARI) score was developed based on angiographic data from the BARI trial.237 It assigns a score to all terminal arteries (terminal portion of the left anterior descending, left circumflex, and right coronary artery, as well as the ramus, diagonals, obtuse marginals, posterior descending and posterolateral branches) based on their length and calibre according to specific criteria. A value of 0 represents an almost insignificant vessel size, whereas a value of 3 defines a large-size artery with a length of two thirds the distance between the base and cardiac apex. Right ventricular marginals and posterior descending artery septal branches are not taken into account. The final score is obtained by dividing the resulting value from the infarctrelated artery by the overall score of all arteries supplying the LV, which finally permits estimation of the percentage of myocardial muscle at risk. An example of the calculation of both the APPROACH and BARI scores for a given patient is presented in Figure 6. The patient had a STEMI due to a proximal left anterior descending artery (LAD) occlusion. Since there was a reperfused diagonal branch of medium importance, the areaat-risk according to the APPROACH score (Figure 5) was 44.5%. For the BARI score, all left and coronary artery branches were classified from 1 to 3, according to their importance. The total value of branches distal to the occluded proximal LAD (total 11 points, corresponding to: distal LAD = 3 points, major diagonal branch = 3 points, 2 minor diagonal branches = 1 point each, 3 septal branches = 1 point each) was divided by the total points of the patient (26), resulting in an area-at-risk of 42.3%. Figure 6 – Area-at-risk calculation according to the APPROACH and BARI scores (adapted from Moral et al234) Methods 89 3.4. TIMI Flow, corrected TFC and TIMI myocardial perfusion grade At the end of the primary PCI procedure, a final run of images of the culprit artery was registered, at 30 frames per second, in order to measure TIMI flow, TIMI frame count and TIMI perfusion grade. If necessary, the view was adjusted, so that the culprit vessel territory was not superimposed. The duration of cine filming was prolonged at least 3 cardiac cycles, to make sure that the entire washout phase was included. These three measures of flow were analysed offline by an operator blinded to other evaluations of the patient. 3.4.1. TIMI flow grade TIMI flow grade was classified according to the standard definition152:  TIMI 3 (complete reperfusion): Anterograde flow into the terminal coronary artery segment through a stenosis is as prompt as anterograde flow into a comparable segment proximal to the stenosis. Contrast material clears as rapidly from the distal segment as from an uninvolved, more proximal segment. If there are difficulties in reproducibly assessing myocardial flow relative to other vessels, the modified definition of TIMI grade 3 flow153 can be used (opacification of the distal coronary artery within three cardiac cycles)  TIMI 2 (partial reperfusion): Contrast material flows through the stenosis to opacify the terminal artery segment. However, contrast enters the terminal segment perceptibly more slowly than more proximal segments. Alternatively, contrast material clears from a segment distal to a stenosis noticeably more slowly than from a comparable segment not preceded by a significant stenosis.  TIMI 1 (penetration with minimal perfusion): A small amount of contrast flows through the stenosis but fails to fully opacify the artery beyond.  TIMI 0 (no perfusion): No contrast flow through the stenosis. 3.4.2. Corrected TIMI frame count The corrected TIMI frame count (cTFC) was measured as the number of frames required for epicardial contrast to reach standardized distal landmarks, as previously described.154 The first frame used for TIMI frame counting was defined as the frame in which a column of dye touched both borders of the coronary artery and moved forward, and the last frame was defined as the frame in which dye begins to enter (but does not necessarily fill) a standard distal landmark in the artery. Methods 96  After cuff release, pulse amplitude increases in the hyperaemic finger. The pulse amplitude recordings are digitized and analysed by an automated, proprietary algorithm. Average pulse amplitude is calculated for each 30-second intervals after cuff occlusion for up to 5 minutes.  As a measure of reactive hyperemia, the pulse amplitude tonometry is then calculated as the ratio of the average amplitude of the PAT signal over a 1-min time interval starting 1 min after cuff deflation divided by the average amplitude of the PAT signal of a 3.5-min time period before cuff inflation (baseline).  Subsequently, PAT index values from the study arm are normalized to the control arm. All these data are analysed by a computer in an operator-independent manner, to get the reactive hyperemia index (RHI) and its logarithmic transformation (L_RHI). In patients with two exams, the same study arm (left or right) was used in both tests. Figure 10 – EndoPAT procedure 3.10. Contrast Enhanced Cardiac Magnetic Resonance Cardiac magnetic resonance was performed in a subset of patients on the 7-8th day post-MI.243, using a 1.5-T magnetic resonance imaging system (Avanto, Siemens Medical System, Erlangen, Germany) equipped with a dedicated cardiac software package and 8 available independent radiofrequency receiver channels, cardiac coil, and vectorcardiogram. Methods 97 After the acquisition of localizing images, longand short-axis cine images were obtained, using retrospectively ECG-gated breath-hold segmented K-space balanced steady-state free precession pulse sequence (trueFISP) technique. The short-axis cine scans of 6-mm slices were used to determine the left ventricular mass, volume, and function (in-plane resolution 1.6x1.6mm; gap 2mm). STIR technique, a triple-IR black-blood turbo spin echo pulse sequence was used for oedema quantification (area at risk). A bolus of contrast medium (gadopentetate dimeglumine - Magnevist, Schering AG, Berlin, Germany) was injected at a dose of 0.2 mmol/kg. Early enhancement images for microvascular obstruction assessment were obtained by acquiring an inversion-recovery segmented gradient echo T1-weighted sequence with a high inversion time (approximately 500 ms), 2-4 min after gadolinium injection. Delayed enhancement images were then obtained by acquiring an inversionrecovery segmented gradient echo T1-weighted sequence, 10 to 15 min after the bolus. All post-processing and analyses of the left ventricular mass, volume, function, area at risk, myocardial infarct size, and presence of microvascular obstruction were performed using CVI 42 Version 5 Software (Circle Cardiovascular Imaging Inc, Calgary, Canada) by a cardiologist experienced in CMR and blinded to all clinical and invasive physiological data. Area-at-risk was manually quantified on short-axis STIR sequences slices, delineating higher intensity areas (no threshold definition) at each slice, with subsequent computation for mass estimation. Infarct size was also assessed manually by planimetry on each short-axis slice, delineating the hyperenhanced area, including areas of hypoenhancement surrounded by the hyperenhanced area, the latter being considered microvascular obstruction. Infarct size, as a percentage of left ventricular mass, was computed from the sum of hyperenhanced pixels from each of the 10 short-axis images divided by the total number of pixels within the left ventricular myocardium multiplied by 100% (21)244. Microvascular obstruction mass was also manually quantified as the sum of hypoenhanced pixels at delayed enhancement sequences as better spatial resolution was found when compared for early enhancement sequences245. 4. STATISTICAL ANALYSIS Continuous variables are presented as mean ± standard deviation (normal distribution) or as median and interquartile range (non-normal distribution); categorical variables are presented as frequencies. All analysis of categorical dependent variables (i.e., RHI<1.67 vs. >1.67, IMR<24 vs >24, etc.) were performed using independent sample T-Test for continuous variables with a normal distribution, Mann-Whitney Test, for continuous variables with a non-normal distribution and Chi-square for categorical variables (with Fisher correction when applicable). Methods 98 Analysis according to RHI tertiles was performed using one-way ANOVA for continuous variables with a normal distribution and Kruskal-Wallis for continuous variables with non-normal distribution. Analysis of IMR or RHI as continuous variables were performed using Pearson’s correlation for continuous variables with a normal distribution and Spearmans’s rho correlation for continuous variables with a non-normal distribution The analysis of IMR predictors was adjusted for confounding variables by linear regression analyses, including the variables identified as relevant on univariate analysis and also all variables considered clinically relevant. For paired comparisons, paired sample T test, Wilcoxon or signs test was used, as indicated. Statistical tests and corresponding p-values were two-sided and a p value <0.05 was considered as statistically significant. IBM SPSS version 21.0 was used for all statistical analyses. 5. ETHICAL ASPECTS The study complied with all ethical international standards, including the World Medical Association’s Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects. It was submitted to and approved by Hospital Prof. Doutor Fernando da Fonseca´s Ethics Committee, Reseaerch Committee and Hospital Administration and to Nova Medical School Ethics Committee. A detailed, written informed consent was obtained from each patient. Inclusion in the study did not affect patient’s care and all patients were treated with the best available resources and knowledge, as usual. 6. FINANTIAL ASPECTS The study was funded by the following companies/foundations:  Astra-Zeneca (unrestricted grant)  Merck Foundation (unrestricted grant)  S. Jude Medical Portugal (pressure-wire offer)  Medtronic (unrestricted grant)  Cordis, Cardinal Health (unrestricted grant)  Bayer (unrestricted grant) 99 RESULTS Results 101 1. Population included Between June 2012 and June 2015, a total of 543 ST elevation myocardial infarction patients treated by primary percutaneous coronary intervention were admitted to Hospital Prof. Doutor Fernando da Fonseca. Of these, 60 patients fulfilled all the inclusion criteria, accepted to participate in the study and were considered eligible for the procedure, according to the operator. The flowchart of patient inclusion is presented in Figure 11Error! Reference source not found.. Figure 11 - Flow diagram of the cohort study Methods 102 1.1. Main epidemiological characteristics Of the 60 patients included, 48 (80.0%) were male. Mean age was 59.6±12.7 years (58.5±12.0 for male and 63.9±14.7 for female patients). The histogram for age is presented in Figure 12. Half of the patients had a previous history of dyslipidaemia and one fourth were diabetics. The main risk factors in the population included are presented in Figure 13. Mean body mass index (BMI) was 27.5±4.0 kg/m2 (28.0±3.8 kg/m2 for males and 25.4±4.2 kg/m2 for females). Mean waist circumference was 99.2±12.2 cm (100.2±12.0 for male and 92.5±12.6 cm for female patients). Figure 12 – Age distribution of patients included Figure 13 – Prevalence of main risk factors Most patients had no history of coronary artery disease: only 9 patients (15.0%) had previous angina complaints and 2 (3.3%) had previously undergone coronary angiography and percutaneous coronary intervention. Accordingly, only 6 patients (10.0%) were on anti-platelet (aspirin or clopidogrel) therapy. Statins were used by 9 patients (15.0%). ACE inhibitors/angiotensin antagonists, beta-blockers and calcium channel antagonists were being used, respectively, by 23 (38.3%), 4 (6.7%) and 4 (6.7%) patients prior to hospital admission. 1.2. Angiographic and angioplasty-related characteristics 1.2.1 Out-of-hospital and in-hospital time delays The median time between the onset of symptoms and the first medical contact (“pain-to-door” time) was 130 minutes (IQR 126 minutes). The median time between the first medical contact and the first balloon dilation (“door-to-balloon” time) was 78 minutes (IQR 45 minutes). The 71,7% 25,0% 43,3% 50,0% Hypertension Diabetes Active smoking habits Dyslipidemia Results 103 median total ischemic time (time between the onset of symptoms and first balloon dilation, or “pain-to-balloon” time) was 209 minutes (IQR 148 minutes). 1.2.2 Angiographic data The culprit artery was the left anterior descending artery (LAD) in 28 patients (46.7%), the left circumflex artery (LCx) in 13 patients (21.7%) and the right coronary artery in 19 patients (31.7%). Multivessel disease (classified as lesions >70% in major coronary arteries) was present in 25 patients (41.7%). The median Syntax score was 15.5 (IQR 10.0). TIMI flow on the first injection was 0 (no flow beyond the point of occlusion) or 1 (faint coronary flow beyond the occlusion with incomplete filling of the distal coronary bed) in 50 patients (83.8%). 1.2.3 Area-at-risk scores Mean APPROACH and BARI scores were similar: 28.5%±6.6% and 28.0%±6.3%, respectively, with a slight tendency for higher areas with the APPROACH score. Correlation between the 2 scores was very high (R=0.90, p<0.001), suggesting that both scores perform similarly well in identifying the area at risk for each lesion location (Figure 14). Figure 14 – Correlation between APPROACH and BARI scores in identifying the area-at-risk for each lesion location Results 104 1.2.4 Angioplasty procedure All patients received intravenous non-fractionated heparin and both aspirin (intravenous) and clopidogrel/ticagrelor (oral). Intracoronary nitrates were also administrated in all patients, since it is mandatory for the IMR measurement. Fourteen patients (23.3%) received intravenous abciximab during angioplasty. Mechanical thrombectomy (aspiration) was performed in 26 cases (43.3%). Stents were implanted in 57 (95.0%) and a direct stenting technique was used in 25 (41.7%), with balloon post-dilation in 28 (38.3%). 1.2.5 Angiographic indicators of microvascular perfusion A normal (TIMI 3) flow was obtained in all patients. Corrected TIMI frame count (cTFC) and TIMI myocardial perfusion grade (TMPG) were measured in all patients immediately after the P-PCI. cTFC median value was 17.0 (IQR 7.0). A normal 3 TMPG was obtained in 37 patients (61.7%). Patients with lower TMPG had, as expected, higher cTFC values (Figure 15). Figure 15 – Corrected TIMI frame count, according to TIMI myocardial perfusion grade (values expressed as median and interquartile range, p value for Kruskal-Wallis test) Results 105 1.3. Index of microcirculatory resistance (IMR) IMR was measured in all patients. Median IMR value was 23.9 (IQR 32.9). An IMR value higher than 40 was measured in 21 patients (35.0%). Two examples of IMR measurements are presented in Figure 16.  Example 1 corresponds to a young 28 years old male patient with an inferior myocardial infarction, due to a proximal RCA occlusion, revascularized (pain-to-balloon time) 155 minutes after symptoms onset. His IMR was 17.  Example 2 corresponds to a 75 years old female patient, with an inferior myocardial infarction revascularized 192 minutes after symptoms onset; she had a proximal LCx occlusion successfully revascularized and her IMR after the procedure was 76. Figure 16 – Examples of IMR measurement Results 112 Table 21 – 2D measurements in first (acute) and second (3 months) echocardiographic exams First (acute) echo Second (3 month) echo p valuec All patients (n=47) Patients w/2 echos (n=45) All patients (n=54) Patients w/ 2 echos (n=45) LVEdV (ml) a 105.8±24.2 105.4±24.8 109.7±26.4 109.8±27.8 0.18 LVEsV (ml) a 54.7±12.7 54.1±12.5 52.9±18.8 53.0±19.6 0.60 LVEF (%) a 47.9±6.7 48.3±6.0 52.6±7.1 52.6±6.7 0.001 WMSI b 1.41 (0.35) 1.41 (0.35) 1.24 (0.35) 1.24 (0.35) <0.001 LA Volume (ml/m2) a 34.8±12.2 34.3±11.8 39.1±15.4 39.8±16.6 0.005 a Presented as mean±standard deviation; b Presented as median (interquartile range); C p-value for the comparison between the first and the second echo only in patients with 2 evaluations; paired samples T-Test for variables with normal distribution and Wilcoxon test for variables with non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. LA – left atria Table 22 – Doppler and 2D speckle tracking measurements in first (acute) and second (3 months) echocardiographic exams First (acute) echo Second (3 month) echo p valueb All patients (n=40) Patients w/2 echos (n=35) All patients (n=51) Patients w/ 2 echos (n=35) Doppler measurements E/A ratio a 1.00±0.34 0.99±0.34 1.20±0.56 1.14±0.51 0.022 E/e’ ratio a 9.00±2.71 9.00±2.84 8.83±3.29 8.89±3.58 0.82 2D speckle tracking imaging GLS a -13.54±2.28 -13.88±2.09 -15.77±3.11 -15.67±3.10 <0.001 a Presented as mean±standard deviation; b p-value for the comparison between the first and the second echo only in patients with 2 evaluations; paired samples T-Test for variables. GLS – global longitudinal strain 1.8. Contrast enhanced cardiac magnetic resonance (ceCMR) Of the 60 patients included in the study:  4 refused to perform the ceCMR, mainly because of claustrophobia;  3 did not perform the exam due to logistic limitations;  2 did the exam, but images were lost;  2 did the exam, but the quality of the images was considered inadequate for evaluation;  49 have the exam available for evaluation of the endpoints defined. Results 113 The exam was performed 8.8 days (6-13) after the primary PCI. Oedema was present in all but 1 patient. Microvascular obstruction was identified in 13 patients (26.5%) and complete transmural necrosis was found in 23 (46.9%). Figure 20 depicts three examples of ceCMR exams, with one case of extensive no-reflow, one case of transmural infarction and one of subendocardial infarction. Figure 20 – Three examples of ceCMR Left column images depict myocardial oedema (area at risk) as assessed by t2-weighted stir (short tau inversion recovery) sequences; the column in the centre shows early enhancement acquisition with a long time of inversion for microvascular obstruction detection; right column shows delayed enhancement sequences for infarct mass estimation. Patient A: large area at risk (between arrows) with extensive no-reflow (yellow arrow), clearly depicted across a transmural antero-septal infarction (short yellow arrow); Patient B: lateral oedema (between arrows) with absence of no-reflow despite the presence of a transmural infarction (*); Patient C: antero-septal oedema (between arrows), negative for the presence of no-reflow and subendocardial myocardial infarction (short yellow arrow) Results 114 2. Outcome measures - summary The primary outcome measures were, as stated above, the IMR values according to the presence of endothelial dysfunction (RHI<1.67) and to RHI tertiles. Since a second EndoPAT measurement was performed, these 2 outcomes will be presented for each EndoPAT exam. Additionally, analysis of RHI as a continuous variable will also be presented for each EndoPAT evaluation. So, in summary, primary outcomes measures will be presented as follows: First EndoPAT 1. IMR values in patients with and without endothelial dysfunction (RHI<1.67) in the first EndoPAT (Section 3.1). 2. IMR values according to RHI tertiles in the first EndoPAT (Section 3.2). 3. Relation of RHI as a continuous variable in the first EndoPAT to IMR values (Section 3.3). Second EndoPAT 4. IMR values in patients with and without endothelial dysfunction (RHI<1.67) in the second EndoPAT (Section 4.1). 5. IMR values according to RHI tertiles in the second EndoPAT (Section 4.2). 6. Relation of RHI as a continuous variable in the second EndoPAT to IMR values (Section 4.3). Secondary outcome measurements included the impact or relation between RHI and IMR and the extent of myocardial necrosis and microvascular reperfusion, evaluated by several different methods. These results will be presented as follows: First EndoPAT 1. Impact of the presence of endothelial dysfunction (RHI<1.67) in the first EndoPAT on the extent of myocardial infarction, evaluated by troponin release, echocardiographic parameters and ceCMR (Section 5.1). 2. Impact of the presence of endothelial dysfunction (RHI<1.67) in the first EndoPAT on microvascular reperfusion indicators, including angiographic indicators (cTFC and TMPG), ECG (ST resolution), and ceCMR (microvascular obstruction) (Section 5.2). Second EndoPAT 3. Impact of the presence of endothelial dysfunction (RHI<1.67) in the second EndoPAT on the extent of myocardial infarction, evaluated by troponin release, echocardiographic parameters and ceCMR (Section 6.1). 4. Impact of the presence of endothelial dysfunction (RHI<1.67) in the second EndoPAT on microvascular reperfusion indicators, including angiographic indicators (cTFC and TMPG), ECG (ST resolution) and ceCMR (microvascular obstruction) (Section 6.2). Results 115 Index of microcirculatory resistance (IMR) 5. Relation between IMR and patient baseline characteristics (Section 7.1) 6. Impact of the presence of coronary microvascular damage as evaluated by an IMR>24 (median value) on the extent of myocardial infarction, evaluated by troponin release, echocardiographic parameters and ceCMR (Section 7.2). 7. Relation between the presence of coronary microvascular damage as evaluated by an IMR>24 (median value) on microvascular reperfusion indicators, including angiographic indicators (cTFC and TMPG), ECG (ST resolution) and ceCMR (microvascular obstruction) (Section 7.3). Results 116 3. Primary outcome – IMR and RHI values on the first EndoPAT 3.1. IMR values in patients with and without endothelial dysfunction (RHI<1.67) on the first EndoPAT Endothelial dysfunction, as evaluated by an RHI<1.67, was present in 11/60 patients (18.3%) in the first EndoPAT evaluation. In the following pages, the main patients characteristics according to the presence of endothelial dysfunction will be shortly described, followed by the IMR analysis. 3.1.1. Main characteristics of patients according to the presence of endothelial dysfunction (RHI<1.67) Table 23 summarizes the main characteristics of patients with and without endothelial dysfunction (RHI<1.67) in the first EndoPAT evaluation. Table 23 – Main characteristics of patients according to the presence of endothelial dysfunction (RHI<1.67) in the first EndoPAT evaluation Variable Total population (n=60) Endothelial dysfunction (RHI<1.67) p value c No (n=49) Yes (n=11) Physical characteristics Age (years) a 59.6±12.7 59.7±13.4 59.1±9.5 0.89 _ Male gender b 48 (80.0) 39 (79.6) 9 (81.8) 0.80 d BMI a 27.5±4.0 27.5±4.0 27.4±4.2 0.94 _ Waist circumference a 99.2±12.2 98.9±12.0 100.8±14.0 0.68 _ Risk Factors and previous coronary disease b Hypertension 43 (71.7) 34 (69.4) 9 (81.8) 0.21 d Diabetes 15 (25.0) 10 (20.4) 5 (45.5) 0.18 d Dyslipidaemia 30 (50.0) 25 (51.0) 5 (45.5) 0.74 _ Active smoking 26 (43.3) 17 (34.7) 9 (81.8) 0.01 d Previous angina 9 (15.0) 7 (14.3) 2 (18.2) 0.89 d Previous revascularization 2 (3.3) 1 (2.0) 1 (9.1) 0.80 d Previous medication b Aspirin 5 (8.3) 3 (6.2) 2 (18.2) 0.50 d Clopidogrel 1 (1.7) 1 (2.1) 0 (0.0) 0.42 d ACEi/ARBs 23 (38.3) 17 (35.4) 6 (54.5) 0.41 d Beta-blockers 4 (6.7) 1 (2.1) 3 (27.3) 0.04 d Nitrates 1 (1.7) 0 (0.0) 0 (0.0) - Statins 9 (15.0) 6 (12.5) 3 (27.3) 0.44 d a Presented as mean±standard deviation; b Presented as number (%); c Independent t-test for continuous variables, Chi-Square for categorical variables; d Yates correction Results 117 Patients whit endothelial dysfunction were more frequently active smokers (81.8% vs. 34.7%, p=0.01), but there were no other significant differences in physical characteristics or in other risk factors between the two groups. Patients with endothelial dysfunction were more likely to be on beta-blocker treatment, but the numbers were too small (1 vs. 3 patients) to allow for any valid conclusion. There were no other significant differences in previous treatment, including anti-platelets and statins. Blood tests on admission, including creatinine, NT-pro-BNP, hsCRP, glucose and HbA1c, were also similar between both populations (Table 24). Table 24 – Laboratory results on admission according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation Variable Total population (n=60) Endothelial dysfunction (RHI<1.67) p value c No (n=49) Yes (n=11) Creatinine (mg/dL) b 0.90 (0.31) 0.90 (0.35) 0.89 (0.16) 0.51 NT-pro-BNP (pg/mL) b 137.5 (255.5) 137.5 (256.3) 191.3 (265.8) 0.99 hsCRP (mg(dL) b 0.44 (0.59) 0.40 (0.58) 0.50 (1.11) 0.46 Glucose (mg/dL) b 133.5 (61.0) 131.0 (60.0) 139.0 (64.0) 0.77 HbA1c (%)b 5.7 (1.1) 5.7 (0.7) 5.7 (1.4) 0.91 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution. 3.1.2. Angiography and angioplasty variables according to the presence of endothelial dysfunction (RHI<1.67) There were no significant differences in ischemic (pain-to-balloon) or hospital-to-balloon times between patients with and without endothelial dysfunction. Main angiographic characteristics, including culprit artery, presence of multivessel disease, Syntax score and initial TIMI flow were also similar in both groups (Table 25). The area-at-risk, measured both by the APPROACH and BARI scores, was also similar. Finally, treatment options (use of mechanical aspiration, stent implantation technique and use of abciximab) were similar in patients with and without RHI<1.67 (Table 26). Results 118 Table 25 – Ischemic times and angiographic characteristics according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation Variable Total population (n=60) Endothelial dysfunction (RHI<1.67) p value d No (n=49) Yes (n=11) Pain-to-balloon time (min)c 209 (148) 210 (156) 174 (128) 0.52 x Door-to-balloon time (min) c 78 (45) 76 (49) 80 (59) 0.37 x Culprit artery b Left anterior descending 28 (46.7) 24 (49.0) 4 (36.4) 0.95 e Left circumflex 13 (21.7) 10 (20.4) 3 (27.3) Right coronary artery 19 (31.7) 15 (30.6) 4 (36.4) Multivessel disease b 25 (41.7) 22 (44.9) 3 (27.3) 0.54 e Syntax score c 15.5 (10.0) 15.5 (11.0) 16.5 (9.0) 0.72 x Area at risk scores a APPROACH score 28.5±6.6 28.1±6.9 30.2±5.5 0.33 x BARI score 27.7±6.3 27.4±6.4 30.6±5.4 0.13 x Initial TIMI flow 0-1 b 50 (83.3) 41 (83.7) 9 (81.8) 0.77 e a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; e Yates correction Table 26 – Angioplasty treatment options according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation Variable a Total population (n=60) Endothelial dysfunction (RHI<1.67) p value b No (n=49) Yes (n=11) Mechanical aspiration 26 (43.3) 21 (42.9) 5 (45.5) 0.86 Balloon pre-dilatation 35 (58.3) 29 (59.2) 6 (54.5) 0.96 Stent implantation 57 (95) 46 (93.9) 11 (100.0) 0.94 Balloon post-dilatation 23 (38.3) 20 (40.5) 3 (27.3) 0.62 Abciximab treatment 14 (23.3) 9 (18.4) 5 (45.5) 0.13 a Presented as number (%); b Chi-Square test, with Yates correction 3.1.3. IMR according to the presence of endothelial dysfunction (RHI<1.67) IMR median values did not differ (and actually tended to be lower) between patients with and without endothelial dysfunction on the first EndoPAT evaluation (Figure 21, Table 27). There was also no difference in the prevalence of microvascular coronary damage, whether this was defined as an IMR >24 (median value) or an IMR >40 (value published in the literature as prognostic marker in STEMI patients). Finally, coronary flow reserve (CFR) values were also similar in both groups (Table 27) Results 119 Table 27 – Invasive hemodynamic measurements according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation Variable a Total Population (n=60) Endothelial Dysfunction (RHI<1.67) p value d No (n=49) Yes (n=11) IMR c 23.9 (32.9) 24.0 (31.2) 16.0 (37.3) 0.17 IMR < 24 (median) b 30 (50) 23 (46.9) 7 (63.6) 0.32 IMR < 40 b 21 (35) 31 (63.9) 8 (72.7) 0.55 Coronary flow reserve c 1.1 (0.8) 1.1 (0.8) 1.2 (2.0) 0.45 Basal SBP a 111.4±28.3 108.7±28.1 122.5±27.6 0.15 Basal DBP a 61.5±13.8 60.3±14.3 66.2±10.3 0.21 Hyperaemic SBP a 91.6±22.5 91.0±23.6 94.2±18.2 0.68 Hyperaemic DBP a 51.3±13.2 50.1±13.2 56.2±12.3 0.63 Aortic pressure a 65.3±15.7 64.1±16.0 69.8±14.2 0.13 Distal pressure a 71.8±23.1 69.5±22.8 81.1±22.8 0.29 a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. Chi-Square for categorical variables Figure 21 – Boxplot of IMR values according to the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation In summary, there were no significant differences in IMR values in patients with and without RHI<1.67 on the first EndoPAT evaluation. With the exception of active smoking and previous beta-blocker therapy, endothelial dysfunction also did not relate to any patient baseline characteristics, including age, risk factors, blood tests on admission, coronary anatomy or procedural aspects of the P-PCI. Results 120 3.2. IMR values according to RHI tertiles on the first EndoPAT According to the defined protocol, RHI was divided in tertiles:  Tertile 1 (n=20): RHI< 1.92  Tertile 2 (n=20): RHI 1.92 – 2.30  Tertile 3 (n=20): RHI> 2.30. In the following pages, the main patient characteristics according to RHI tertiles will be shortly described, followed by the IMR analysis in these groups. 3.2.1. Main characteristics of patients according to tertiles of RHI Table 28 summarizes the main characteristics of patients according to RHI tertiles on the first EndoPAT evaluation. There were no significant differences in any of the analysed variables, including physical characteristics, risk factors and previous pharmacological treatment. Blood tests results on admission were also similar (Table 29). Table 28 – Main characteristics of patients according to RHI tertiles on the first EndoPAT evaluation Variable Total population (n=60) RHI p valuec Tertile 1 (n=20) Tertile 2 (n=20) Tertile 3 (n=20) Physical characteristics Age (years) a 59.6±12.7 58.9±11.6 58.8±15.1 61.0±11.4 0.83 x Male gender b 48 (80.0) 16 (80.0) 15 (75.0) 17 (85.0) 0.89 d BMI a 27.5±4.0 27.6±3.93 28.4±4.02 26.5±3.99 0.33 x Waist circumference a 99.2±12.2 100.4±14.1 102.4±13.4 99.2±12.2 0.25 x Risk factors and previous coronary disease b Hypertension 43 (71.7) 16 (80.0) 13 (65.0) 14 (70.0) 0.56 x Diabetes 15 (25.0) 6 (30.0) 5 (25.0) 4 (20.0) 0.77 x Dyslipidaemia 30 (50.0) 9 (45.0) 10 (50.0) 11 (55.0) 0.82 x Active smoking 26 (43.3) 12 (60.0) 7 (35.0) 7 (35.0) 0.18 x Previous angina 9 (15.0) 4 (20.0) 2 (10.0) 3 (15.0) 0.86 d Previous revascularization 2 (3.3) 1 (5.0) 1 (5.0) 0 (0.0) 0.94 d Previous medication b Aspirin 5 (8.3) 2 (10.0) 1 (5.0) 2 (10.0) 0.97 d ACEi/ARBs 23 (38.3) 9 (45.0) 7 (35.0) 7 (35.0) 0.79 x Beta-blockers 4 (6.7) 3 (15.0) 1 (5.0) 0 (0.0) 0.44 d Statins 9 (15.0) 4 (20.0) 3 (15.0) 2 (10.0) 0.88 d a Presented as mean±standard deviation; b Presented as number (%); c One-way ANOVA for continuous variables, Chi-Square for categorical variables; d Yates correction Results 121 Table 29 – Laboratory results on admission according to RHI tertiles on the first EndoPAT evaluation Variable Total Population (n=60) RHI p valuec Tertile 1 (n=20) Tertile 2 (n=20) Tertile 3 (n=20) Creatinine (mg/dL) b 0.90 (0.31) 0.90 (0.17) 0.90 (0.40) 0.96 (0.36) 0.62 NT-pro-BNP (pg/mL) b 137.5 (255.5) 162.0 (269.0) 113.0 (170.5) 131.5 (298.0) 0.64 hs-CRP (mg(dL) b 0.44 (0.59) 0.64 (0.71) 0.29 (0.35) 0.44 (0.95) 0.10 Glucose (mg/dL) b 133.5 (61.0) 122.0 (54.0) 141.5 (60.0) 131.0 (61.0) 0.63 HbA1c (%)b 5.7 (1.1) 5.5 (0.9) 5.8 (1.9) 5.7 (2.3) 0.44 a Presented as mean±standard deviation; b Presented as median (interquartile range); c One-way ANOVA for continuous variables with a normal distribution, Kruskal-Wallis test for continuous variables with a non-normal distribution. 3.2.2. Angiography and angioplasty variables according to RHI tertiles There were no significant differences in pain-to-balloon or hospital-to-balloon times in RHI tertiles groups. Main angiographic characteristics, including culprit artery, presence of multivessel disease, Syntax score and initial TIMI flow were also similar (Table 30). Table 30 – Ischemic times and angiographic characteristics according RHI tertiles on the first EndoPAT evaluation Variable Total Population (n=60) RHI p valued Tertile 1 (n=20) Tertile 2 (n=20) Tertile 3 (n=20) Pain-to-balloon time (min)c 209 (148) 223 (149) 201 (179) 200 (102) 0.85 x Door-to-balloon time (min)c 78 (45) 82 (45) 67 (55) 76 (43) 0.43 x Culprit artery b Left anterior descending 28 (46.7) 10 (50.0) 8 (40.0) 10 (50.0) 0.99 e Left circumflex 13 (21.7) 4 (20.0) 5 (25.0) 4 (20.0) Right coronary artery 19 (31.7) 6 (30.0) 7 (35.0) 6 (30.0) Multivessel disease b 25 (41.7) 11 (55.0) 8 (40.0) 6 (30.0) 0.28 x Syntax score c 15.5 (10.0) 16.0 (9.0) 11.8 (12.0) 15.8 (7.0) 0.95 x Area at risk scores APPROACH score a 27.8 (2.0) 27.8 (4.0) 27.8 (4.0) 27.8 (2.0) 0.63 x BARI score a 27.7±6.3 29.3±7.3 28.0±5.7 26.7±6.0 0.46 x Initial TIMI flow 0-1 b 50 (83.3) 17 (85.0) 19 (95.0) 14 (70.0) 0.23 e a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d One-way ANOVA for continuous variables with a normal distribution, Kruskal-Wallis test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; e Yates correction The area-at-risk, measured both by the APPROACH and BARI scores, was also similar in the three tertiles. Finally, treatment options (use of mechanical aspiration, stent implantation technique and use of abciximab) was again, similar (Table 31). Results 128 Table 41 – Ischemic times and angiographic characteristics according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Variable Total population (n=38) Endothelial dysfunction (RHI<1.67) p value d No (n=22) Yes (n=16) Pain-to-balloon time (min)c 209 (173) 209 (186) 211 (167) 0.94 Door-to-balloon time (min) c 75 (52) 79 (46) 57 (44) 0.06 Culprit artery b Left anterior descending 21 (53.3) 11 (50.0) 10 (62.5) 0.89 f Left circumflex 7 (18.4) 5 (22.7) 2 (12.5) Right coronary artery 10 (26.3) 6 (27.3) 4 (25.0) Multivessel disease b 19 (50.0) 8 (36.4) 11 (68.8) 0.10 f SYNTAX score c 17.8±6.2 16.8±6.2 19.0±6.0 0.29 Area at risk scores a APPROACH score 27.8 (3.0) 28.1 (5.0) 29.7 (10.0) 0.08 BARI score 28.5 (6.0) 26.7 (7.0) 30.2 (10.0) 0.07 Initial TIMI flow 0-1 b 31 (81.6) 16 (72.7) 15 (93.8) 0.22 a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; E Yates correction Table 42 – Angioplasty treatment options according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Variable a Total population (n=38) Endothelial dysfunction (RHI<1.67) p value b No (n=22) Yes (n=16) Mechanical aspiration 16 (42.1) 9 (40.9) 7 (43.8) 0.86 Balloon pre-dilatation 21 (55.3) 13 (59.1) 8 (50.0) 0.58 Stent implantation 36 (94.7) 20 (90.9) 16 (100.0) 0.61 Balloon post-dilatation 24 (63.2) 14 (63.6) 10 (62.5) 0.94 Abciximab treatment 9 (23.7) 4 (18.2) 5 (31.5) 0.58 Mechanical aspiration 16 (42.1) 9 (40.9) 7 (43.8) 0.86 a Presented as number (%); b Chi-Square test, with Yates correction 4.1.3. IMR according to the presence of endothelial dysfunction (RHI<1.67) There was a clear trend towards higher values of IMR in patients with endothelial dysfunction on the second EndoPAT evaluation: median values (IQR) 40.5 (54.4) vs. 22.0 (26.0) in patients without endothelial dysfunction (p=0.09) (Figure 22, Table 43). The prevalence of microvascular coronary damage, either defined as an IMR >24 (median value) or an IMR >40 (value published in the literature as prognostic marker in STEMI patients), was almost 2 times higher in patients with endothelial dysfunction, although differences did not reach statistical significance (Table 43). Other invasive variables, including coronary flow reserve (CFR) values, were similar in both groups. Results 129 Table 43 – Invasive hemodynamic measurements according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Variable a Total population (n=38) Endothelial dysfunction (RHI<1.67) p value d No (n=22) Yes (n=16) IMR c 23.4 (35.2) 22.0 (26.0) 40.5 (54.4) 0.09 IMR < 24 (median) b 18 (47.4) 8 (36.4) 10 (62.5) 0.11 IMR < 40 b 14 (36.8) 6 (27.3) 8 (50.0) 0.15 Coronary flow reserve c 1.00 (0.70) 1.10 (0.60) 1.00 (0.90) 0.92 Basal SBP a 112.0±31.3 113.0±26.3 110.4±38.8 0.81 Basal DBP a 61.3±14.0 59.0±14.0 65.0±13.7 0.22 Hyperaemic SBP a 92.8±22.0 96.1±19.3 87.8±25.4 0.28 Hyperaemic DBP a 51.7±12.1 52.7±11.1 50.1±13.8 0.56 Aortic pressure a 75.7±25.1 74.7±20.3 77.2±31.7 0.72 Distal pressure a 64.6±14.9 67.1±14.3 60.9±15.5 0.24 a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. Chi-Square for categorical variables. Figure 22 – Boxplot of IMR values according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation In summary, the presence of endothelial dysfunction (defined as an RHI<1.67) on the second EndoPAT was associated with a trend for higher IMR values measured immediately after P-PCI. These patients also tended to have more severe coronary artery disease, higher areas-at-risk and worse initial TIMI flow. Results 130 4.2. IMR values according to RHI tertiles on the second EndoPAT According to the defined protocol, RHI in the second endothelial dysfunction was divided in tertiles:  Tertile 1 (n=13): RHI< 1.62  Tertile 2 (n=13): RHI 1.62 – 1.96  Tertile 3 (n=12): RHI> 1.96 . In the following pages, main patient characteristics according to RHI tertiles will be shortly described, followed by the IMR analysis in these groups. 4.2.1. Main characteristics of patients according to tertiles of RHI Table 44 summarizes the main characteristics of patients according to RHI tertiles on the second EndoPAT evaluation. There was a trend towards more male patients and more previous use of ACEi/ARBs for higher tertiles of RHI. There were no significant differences in any of the other variables, including physical characteristics, risk factors and other previous pharmacological treatment. Blood tests on admission (Table 45) were also similar. Table 44 – Main characteristics of patients according to RHI tertiles on the second EndoPAT evaluation Variable Total population (n=38) RHI p valuec Tertile 1 (n=13) Tertile 2 (n=13) Tertile 3 (n=12) Physical characteristics Age (years) a 60.0±13.7 62.8±14.4 54.8±13.6 62.6±12.3 0.25 x Male gender b 29 (76.3) 8 (61.5) 10 (76.9) 11 (91.7) 0.41 d BMI a 27.2±4.0 26.3±3.2 28.9±3.9 26.4±4.5 0.17 x Waist circumference a 100.0±11.5 96.8±10.6 102.8±12.9 100.0±10.8 0.50 x Risk Factors and previous coronary disease b Hypertension 26 (68.4) 8 (61.5) 9 (69.2) 9 (75.0) 0.93 d Diabetes 12 (31.6) 3 (23.1) 7 (53.8) 2 (16.7) 0.24 d Dyslipidaemia 19 (50.0) 7 (53.8) 5 (38.5) 7 (58.3) 0.58 x Active smoking 13 (34.2) 3 (23.1) 6 (46.2) 4 (33.3) 0.69 d Previous angina 7 (18.4) 1 (7.7) 2 (15.4) 4 (33.0) 0.51 d Previous revascularization 2 (5.3) 1 (7.7) 0 (0.0) 1 (8.3) 0.94 d Previous medication b Aspirin 3 (8.1) 1 (7.7) 0 (0.0) 2 (16.7) 0.64 d ACEi/ARBs 13 (35.1) 2 (15.4) 5 (38.5) 6 (50.0) 0.43 d Beta-blockers 1 (2.7) 0 (0.0) 0 (0.0) 1 (8.3) 0.88 d Statins 4 (10.8) 1 (7.7) 3 (23.1) 0 (0.0) 0.46 d a Presented as mean±standard deviation; b Presented as number (%); c One-way ANOVA for continuous variables, Chi-Square for categorical variables; d Yates correction Results 131 Table 45 –Laboratory results on admission according to RHI tertiles on the second EndoPAT evaluation Variable Total population (n=38) RHI p valuec Tertile 1 (n=13) Tertile 2 (n=13) Tertile 3 (n=12) Creatinine (mg/dL) b 0.91±0.28 0.87±0.16 0.95±0.32 0.90±0.34 0.72 NT-pro-BNP (pg/mL) b 158.5 (305) 163.0 (204.0) 299.0 (290.0) 103.0 (434.8) 0.94 hs-CRP (mg(dL) b 0.57 (0.71) 0.60 (0.99) 0.95 (0.72) 0.29 (0.14) 0.12 Glucose (mg/dL) b 136.0 (63.0) 136.0 (108.5) 169.0 (132.5) 122.0 (43.8) 0.29 HbA1c (%)b 6.6 (1.3) 5.8 (0.9) 6.3 (4.3) 5.7 (0.6) 0.37 a Presented as mean±standard deviation; b Presented as median (interquartile range); c One-way ANOVA for continuous variables with a normal distribution, Kruskal-Wallis test for continuous variables with a non-normal distribution. 4.2.2. Angiography and angioplasty variables according to RHI tertiles There were no significant differences in pain-to-balloon or hospital-to-balloon times in RHI tertiles groups. There was a clear trend towards more complex coronary disease in lower tertiles of RHI: multivessel disease was present in 75.0% of patients in the lower tertile, as compared to 30.8% in patients in the upper tertile. Syntax score also tended to be higher in lower tertiles of RHI (Table 46). Table 46 – Ischemic times and angiographic characteristics according RHI tertiles on the second EndoPAT evaluation Variable Total population (n=38) RHI p valued Tertile 1 (n=13) Tertile 2 (n=13) Tertile 3 (n=12) Pain-to-balloon time (min)c 209 (173) 257 (173) 235 (185) 247 (155) 0.69 Door-to-balloon time (min)c 75 (52) 55 (44) 63 (51) 82 (45) 0.52 Culprit artery b Left anterior descending 21 (55.5) 9 (69.2) 8 (61.5) 4 (33.3) 0.61 e Left circumflex 7 (18.4) 1 (7.7) 3 (23.1) 3 (25.0) Right coronary 10 (26.3) 3 (23.1) 2 (15.4) 5 (41.7) Multivessel disease b 19 (50.0) 9 (75.0) 6 (46.2) 4 (30.8) 0.08 Syntax score c 17.8±6.1 20.4±5.7 17.0±6.0 15.8±6.4 0.15 Area at risk scores APPROACH score c 27.8 (3.0) 29.7 (10.0) 29.7 (10.0) 27.7 (8.0) 0.07 BARI score a 29.1±6.7 31.1±7.4 31.4±5.3 24.5±5.1 0.011 Initial TIMI flow 0-1 b 31 (81.6) 12 (92.3) 10 (76.9) 9 (75.0) 0.79 e a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d One-way ANOVA for continuous variables with a normal distribution, Kruskal-Wallis test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; e Yates correction Results 132 The area-at-at risk evaluated by the BARI score was significantly lower in patients in the third tertile (higher values) of RHI and a similar trend was found when the APPROACH score was used (p=0.07). Initial unfavourable TIMI flow (0-1) was also more prevalent in the lowest RHI tertile, although the difference was not statistically significant (Table 46). Finally, treatment options (use of mechanical aspiration, stent implantation technique or use of abciximab) was similar between the 3 groups (Table 47). Table 47 – Angioplasty treatment options according to RHI tertiles on the second EndoPAT evaluation Variable a Total population (n=38) RHI p value b Tertile 1 (n=13) Tertile 2 (n=13) Tertile 3 (n=12) Mechanical aspiration 16 (42.1) 5 (38.5) 5 (38.5) 6 (50.0) 0.80 x Balloon pre-dilatation 21 (55.3) 7 (53.8) 6 (46.2) 8 (66.7) 0.58 x Stent implantation 36 (94.7) 13 (100.0) 12 (92.3) 11 (91.7) 0.94 c Balloon post-dilatation 14 (36.8) 4 (30.8) 7 (53.8) 3 (25.0) 0.52 c Abciximab 9 (23.7) 5 (38.5) 2 (15.4) 2 (16.7) 0.59 c a Presented as number (%); b Chi-Square test; c Yates correction 4.2.3. IMR according to RHI tertiles IMR median values were lower in the third tertile of RHI, although this difference was not statistically significant. The number of patients with coronary microvascular dysfunction (IMR>24 or IMR>40) also decreased from the first to the third tertile of RHI. CRF were similar between groups (Table 48). Table 48 – Invasive hemodynamic measurements according to RHI tertiles on the second EndoPAT evaluation Variable Total population (n=38) RHI p value c Tertile 1 (n=13) Tertile 2 (n=13) Tertile 3 (n=12) IMR a 23.4 (35.2) 39.0 (43.4) 23.8 (42.5) 19.5 (30.6) 0.64 x IMR < 24 (median) b 18 (47.4) 8 (61.5) 6 (46.2) 4 (33.3) 0.37 x IMR < 40 b 14 (36.8) 6 (46.2) 5 (38.5) 3 (25.0) 0.78 c Coronary flow reserve a 1.1 (0.8) 1.3 (0.84) 1.0 (0.50) 1.1 (0.9) 0.36 x a Presented as median (interquartile range); b Presented as number (%); c Mann-Whitney test for continuous variables with a nonnormal distribution. Chi-Square for categorical variables In summary, there was a trend for lower IMR values in third tertile of RHI measured on the second EndoPAT. These patients also had lower areas-at-risk and a trend to less complex coronary artery disease (lower Syntax and lower prevalence of multivessel disease) Results 133 4.3. Relation between RHI as a continuous variable on the second EndoPAT and IMR values 4.3.1. RHI as a continuous variable and main baseline patient characteristics Male patients had significantly higher RHI values on the second EndoPAT evaluation (Table 49). Table 49 – Correlations between RHI on the second EndoPAT and baseline continuous variables Variable Correlation (R) p value a Physical characteristics Age -0.006 0.970 BMI 0.033 0.970 Waist circumference 0.132 0.470 Admission laboratory parameters Creatinine (mg/dl) 0.030 0.258 NT-pro-BNP (pg/mL) -0.193 0.259 hs-CRP (mg/dL) 0.249 0.131 Glucose (mg/dL) -0.107 0.528 HbA1c (%) 0.008 0.965 a Spearmans’s rho Table 50 – RHI values on the second EndoPAT according to baseline categorical variables Variable Variable present? a p value b No Yes Physical characteristics Male gender (29/38) 1.43±0.53 2.00±0.56 0.024 Risk factors and previous coronary disease Hypertension (26/38) 1.84±0.46 1.88±0.67 0.87 Diabetes (12/38) 1.92±0.68 1.76±0.38 0.46 Dyslipidaemia (19/38) 1.84±0.65 1.89±0.56 0.78 Active smoking (13/38) 1.86±0.70 1.87±0.37 0.99 Previous angina (7/38) 1.84±0.64 1.98±0.41 0.58 Previous revascularization (2/38) 1.89±0.59 1.50±0.99 0.38 Previous medication Aspirin (3/38) 1.85±0.53 2.16±1.35 0.40 ACEi/ARBs (13/38) 1.72±0.49 2.16±0.71 0.042 Statins (9/38) 1.71±0.10 1.89±0.64 0.57 a RHI presented as mean±standard deviation; b Independent t-test for continuous variables with a normal distribution. MannWhitney test for continuous variables with a non-normal distribution Results 134 No significant associations were found between RHI and any other baseline patient characteristics, including age, physical characteristics, risk factors and blood tests on admission (Table 49, Table 50). However, diabetic patients and patients with previous coronary revascularization tended to have lower RHI values. Patients treated previously with ACEi/ARBs had significantly higher RHI values, and those treated previously with statins and aspirin also showed a trend for higher RHI values (Table 50). 4.3.2. RHI as a continuous variable and angiography/angioplasty variables There were no significant associations between RHI measured on the second EndoPAT and ischemic times. RHI tended to be lower in patients with multivessel disease, lower initial TIMI flow and LAD as the culprit artery (Table 51, Table 52). Table 51 – Correlations between RHI on the second EndoPAT and angiography and P-PCI related continuous variables Variable Correlation (R) p value a Pain-to-balloon time (min) -0.107 0.523 Door-to-balloon time (min) 0.263 0.111 Syntax score -0.241 0.144 Area at risk scores APPROACH score -0.426 0.008 BARI score -0.361 0.026 a Pearson’s Correlation for continuous variables with a normal distribution; Spearmans’s rho for continuous variables without a normal distribution Table 52 – RHI values on the second EndoPAT according to angiography and P-PCI categorical variables Variable Variable present? a p value b No Yes Angiography Culprit artery = LAD (21/38) 2.02±0.51 1.74±0.65 0.15 Multivessel disease (20/38) 2.00±0.61 1.71±0.58 0.13 Initial TIMI flow 0-1 (29/38) 2.05±0.43 1.81±0.64 0.30 Primary PCI Mechanical aspiration (16/38) 1.79±0.61 1.96±0.60 0.40 Balloon pre-dilatation (21/38) 1.84±0.52 1.89±0.68 0.79 Stent implantation (36/38) 2.71±1.24 1.82±0.54 0.21 Balloon post-dilatation (14/38) 1.80±0.58 1.98±0.65 0.39 Abciximab use (9/38) 1.90±0.68 1.76±0.25 0.57 a RHI presented as mean±standard deviation; b Independent t-test for continuous variables with a normal distribution. MannWhitney test for continuous variables with a non-normal distribution Results 135 There was a significant negative correlation between both area-at-risk-scores and RHI, suggesting that patients with higher areas at risk had lower RHI values on the second EndoPAT evaluation (Table 51). 4.3.3. RHI as a continuous variable and IMR There was no relation between RHI measured on the second EndoPAT and IMR or any of the invasive measurements performed (Table 53). However, there was a trend for lower RHI values in patients with coronary microvascular dysfunction, according to the IMR cut-offs of 24 and 40 (Table 54). Table 53 – RHI on the second EndoPAT and invasive hemodynamic continuous variables Variable Correlation (R) p value a IMR -0.090 0.593 Coronary flow reserve 0.018 0.917 Basal SBP 0.016 0.925 Basal DBP -0.227 0.184 Hyperaemic SBP 0.200 0.250 Hyperaemic DBP 0.037 0.833 Aortic pressure 0.181 0.298 Distal pressure -0.030 0.862 a Pearson’s Correlation for continuous variables with a normal distribution; Spearmans’s rho for continuous variables without a normal distribution Table 54 – RHI values on the second EndoPAT according to IMR thresholds Variable Variable present? p value b No Yes IMR < 24 (median) (18/38) 1.92±0.59 1.81±0.62 0.56 IMR < 40 (14/38) 1.92±0.68 1.77±0.45 0.48 a Presented as mean±standard deviation; b Mann-Whitney test for continuous variables with a non-normal distribution In summary, there was a trend for lower RHI values in patients with coronary microvascular dysfunction (increased IMR). RHI values correlated with male gender and previous treatment with ACEi/ARBs and tended to be lower in diabetic and previously revascularized patients. Additionally, RHI values on the second EndoPAT correlated with the area-at-risk and tended to be lower in patients with more complex coronary artery disease) Results 136 5. Secondary outcome – Extent of myocardial infarction and microvascular reperfusion according to RHI on the first EndoPAT 5.1. Impact of the presence of endothelial dysfunction (RHI<1.67) on the first EndoPAT on the extent of myocardial infarction The extent of infarction was evaluated by:  The area under the curve and the peak values of 7 evaluations of troponin I in the first 48 hours after the P-PCI,  The area of infarction in the contrast enhanced cardiac magnetic resonance,  And, indirectly, by echocardiography parameters, including left ventricular volumes and ejection fraction, wall motion score index and global longitudinal strain, both on the initial and follow-up (3 month) echocardiograms. 5.1.1. Endothelial dysfunction and troponin release The peak values and the area under the curve of the 7 values of troponin I in the first 48 hours, are presented in Table 55. Both total values and values indexed to the area-at-risk scores (APPROACH and BARI) are shown, in patients with and without endothelial dysfunction (RHI<1.67) on the first EndoPAT evaluation. There were no significant differences in these populations. Table 55 – Impact of endothelial dysfunction (RHI<1.67, measured on first EndoPAT) on Troponin I release Variable a, b Total Population (n=60) Endothelial Dysfunction (RHI<1.67) p value c No (n=49) Yes (n=11) TnIpeak 117±82 117±87 115±55 0.96 TnIpeak (APPROACH) 34±27 33±29 35±19 0.82 TnIpeak (BARI) 33±25 32±26 36±20 0.62 TnIpeak (2 scores) 33±26 33±27 36±19 0.72 TnIAUC 1938±1283 1951±1376 1883±787 0.88 TnIAUC (APPROACH) 565±441 564±472 574±275 0.94 TnIAUC (BARI) 548±405 541±43 584±282 0.75 TnIAUC (2 scores) 557±421 552±449 579±278 0.85 a Presented as mean±standard deviation; b Peak value and area under the curve (AUC) of 7 troponin I (TnI, in mg/dL) measurements performed in the first 48 hours after the primary angioplasty; total values and values indexed to the APPROACH, BARI or both are presented; c Independent t-test. Results 137 5.1.2. Endothelial dysfunction and echocardiography parameters There were no significant differences in 2D, Doppler or 2D speckle tracking imaging measurements in the first echocardiogram between patients with and without endothelial dysfunction on the first EndoPAT (Table 56). Likewise, all measurements were similar in the second echocardiogram in patients with and without RHI<1.67 (Table 57). Table 56 – Impact of endothelial dysfunction (RHI<1.67, measured on first EndoPAT) on the first echocardiogram parameters Variable a, b Total population Endothelial dysfunction (RHI<1.67) p value c No Yes 2D measurements (n=60) (n=49) (n=11) LVEdV (ml) a 105.8±24.2 107.5±24.0 95.7±24.4 0.27 LVEsV (ml) a 54.7±12.7 55.6±12.9 49.5±11.1 0.28 LVEF (%)a 47.9±6.7 47.9±6.8 47.8±5.9 0.96 Wall motion score index b 1.41 (0.35) 1.41 (0.35) 1.53 (0.18) 0.78 Left atria volume (ml/m2) a 34.8±12.2 35.7±12.9 30.6±7.2 0.32 Doppler measurements (n=40) (n=34) (n=7) E/A ratio a 1.00±0.34 0.99±0.34 1.06±0.36 0.68 E/e’ ratio a 9.00±2.71 8.91±2.85 9.58±1.55 0.61 2D speckle tracking imaging (n=40) (n=34) (n=7) Global longitudinal strain a -13.54±2.28 -13.38±2.30 -14.7±1.96 0.23 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. Table 57 – Impact of endothelial dysfunction (RHI<1.67, measured on first EndoPAT) on the 3 month echocardiogram parameters Variable a, b Total population Endothelial dysfunction (RHI<1.67) p value c No Yes 2D measurements (n=54) (n=43) (n=11) LVEdV (ml) a 109.7±26.4 111.1±28.5 103.8±13.9 0.44 LVEsV (ml) a 52.9±18.8 54.2±20.4 47.6±8.8 0.32 LVEF (%)a 52.6±7.1 52.2±7.4 54.2±5.2 0.43 Wall motion score index b 1.24 (0.35) 1.24 (0.35) 1.29 (0.24) 0.93 Left atria volume (ml/m2) a 39.1±15.4 40.4±16.7 34.3±7.2 0.25 Doppler measurements (n=51) (n=40) (n=11) E/A ratio a 1.20±0.56 1.21±0.54 1.15±0.66 0.75 E/e’ ratio a 8.92±3.29 8.83±3.51 9.29±2.30 0.69 2D speckle tracking imaging (n=40) (n=40) (n=11) Global longitudinal strain a -15.77±3.11 -15.80±3.36 -15.68±2.11 0.91 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. Results 144 6. Secondary outcome – Extent of myocardial infarction and microvascular reperfusion according to RHI on the second EndoPAT 6.1. Impact of the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT on the extent of myocardial infarction 6.1.1. Endothelial dysfunction and Troponin release The peak values and the area under the curve of troponin I, according to the presence of endothelial dysfunction on the second EndoPAT evaluation are presented in Table 63. Both total values and values indexed to the area-at-risk scores (APPROACH and BARI) are presented, in patients with and without endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation. Patients with endothelial dysfunction had significantly higher values of both peak TnI and AUC of TnI, both total and indexed to the area-at-risk scores. Table 63 – Impact of endothelial dysfunction (RHI<1.67, measured on second EndoPAT) on Troponin I release Variable a, b Total population (n=38) Endothelial dysfunction (RHI<1.67) p value c No (n=22) Yes (n=16) TnIpeak 95 (96) 67 (81) 118 (186) 0.024 TnIpeak (APPROACH) 25 (32) 17 (20) 34 (56) 0.009 TnIpeak (BARI) 24 (31) 17 (22) 33 (47) 0.008 TnIpeak (2 scores) 24 (31) 17 (21) 33 (55) 0.009 TnIAUC 1293 (1580) 1076 (1042) 2305 (2486) 0.012 TnIAUC (APPROACH) 403 (522) 315 (303) 664 (1080) 0.008 TnIAUC (BARI) 383 (448) 314 (326) 618 (799) 0.007 TnIAUC (2 scores) 393 (482) 314 (300) 641 (984) 0.007 a Presented as median (IQR); b Peak value and area under the curve (AUC) of 7 troponin I (TnI, in mg/dL) measurements performed in the first 48 hours after the primary angioplasty; total values and values indexed to the APPROACH, BARI or both are presented; c Mann-Whitney Test. 6.1.2. Endothelial dysfunction and echocardiography parameters Patients with endothelial dysfunction had significantly higher end-systolic volumes, lower LVEF and worse wall motion score index in the first echocardiogram. Accordingly with the difference Results 145 in WSMI, they also had higher values of global longitudinal strain (Table 64). These differences were no longer visible in the echocardiogram performed at 3 month (Table 65). Table 64 – Impact of endothelial dysfunction (RHI<1.67, measured on second EndoPAT) on the first echocardiogram parameters Variable Total population Endothelial dysfunction (RHI<1.67) p value c No Yes 2D measurements (n=26) (n=15) (n=11) LVEdV (ml) a 109.8±23.4 106.5±30.2 113.8±11.4 0.48 LVEsV (ml) a 55.9±12.6 51.1±12.3 61.7±10.9 0.047 LVEF (%)a 48.6±7.1 51.4±4.7 45.3±8.3 0.045 Wall motion score index b 1.44 (0.41) 1.35 (0.47) 1.77 (0.47) 0.006 Left atria volume (ml/m2) a 36.1±11.0 36.7±12.1 35.3±10.0 0.78 Doppler measurements (n=21) (n=11) (n=10) E/A ratio a 1.10±0.40 1.17±0.40 0.97±0.0.39 0.30 E/e’ ratio a 8.91±3.30 8.83±3.91 9.05±2.18 0.89 2D speckle tracking imaging (n=21) (n=11) (n=10) Global longitudinal strain a -13.16±2.35 -14.32±1.72 -11.89±2.35 0.014 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. Table 65 – Impact of endothelial dysfunction (RHI<1.67, measured on second EndoPAT) on the 3 month Echocardiogram parameters Variable Total Population Endothelial Dysfunction (RHI<1.67) p value c No Yes 2D measurements (n=30) (n=19) (n=11) LVEdV (ml) a 109.7±28.1 103.9±29.8 119.7±22.4 0.14 LVEsV (ml) a 53.8±21.4 49.8±21.0 60.7±21.2 0.18 LVEF (%)a 52.2±7.9 53.2±7.6 50.5±8.3 0.37 Wall motion score index b 1.24 (0.40) 1.21 (0.35) 1.29 (0.66) 0.33 Left atria volume (ml/m2) a 40.6±15.4 42.2±16.9 38.0±12.7 0.46 Doppler measurements (n=29) (n=19) (n=10) E/A ratio a 1.31±0.68 1.30±0.62 1.34±0.82 0.87 E/e’ ratio a 8.96±3.93 9.13±4.19 8.59±3.51 0.74 2D speckle tracking imaging (n=29) (n=19) (n=10) Global longitudinal strain a -15.43±3.55 -15.95±3.01 -14.4±4.39 0.28 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. When the 2 echocardiograms were compared, significant improvements in the wall motion score index were found in both patients with and without endothelial dysfunction. There was also a trend for an improvement in the global longitudinal strain in the group without Results 146 endothelial dysfunction. There were no other significant differences, independently of the presence of endothelial dysfunction. Table 66 – Baseline and 3 months echocardiographic parameters according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation RHI > 1.67 RHI < 1.67 Echo parameters Echo1 Echo2 P valuec Echo1 Echo2 P valuec 2D measurements n=15 n=11 LVEdV (ml) a 106.5±3..2 101.0±33.0 0.33 113.8±11.8 119.7±22.4 0.61 LVEsV (ml) a 51.1±12.3 48.3±22.3 0.24 61.7±10.9 60.7±21.1 1.00 LVEF (%)a 51.4±4.8 53.2±5.1 0.24 45.3±8.3 50.5±8.3 0.24 WMSI b 1.35 (0.47) 1.18 (0.35) 0.012 1.65 (0.47) 1.24 (0.74) 0.049 Left atria (ml/m2) a 36.7±12.1 43.8±18.8 0.16 35.3±10.0 38.0±12.7 0.16 Doppler measurements n=12 n=11 E/A' ratio a 1.14±0.40 1.24±0.43 0.61 0.97±0.39 1.34±0.82 0.08 E/e’ ratio a 9.0±4.1 8.2±4.7 0.37 9.1±2.2 8.6±3.5 0.72 2D speckle tracking imaging n=10 n=10 Global longitudinal strain a -14.3±1.8 -16.1±2.6 0.07 -11.9±2.3 -14.4±4.4 0.53 a Presented as mean±standard deviation; b Presented as median (interquartile range); C p-value for the comparison between the first and the second Echo only in patients with 2 evaluations; paired samples T-Test for variables with normal distribution and Wilcoxon test for variables with non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. 6.1.3. Endothelial dysfunction and contrast enhanced cardiac magnetic resonance Contrast enhanced CMR was performed in 29 of the 38 patients with a second EndoPAT evaluation. Of these, 11 (37.9%) had endothelial dysfunction (RHI<1.67) (Table 67). Patients with endothelial dysfunction (RHI<1.67) had significantly lower left ventricular ejection fraction and higher wall motion score index, as compared to patients with RHI>1.67. There was a trend for more transmural necrosis (22.2% vs. 63.6%, p=0.06) and higher infarct mass (median value 10.1 vs. 17.5, p=0.08) in patients with RHI<1.67. The percent mass of infarct indexed to area-at-risk scores (APPROACH and BARI) also tended to be higher in patients with endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation. Results 147 Table 67 – Impact of endothelial dysfunction (RHI<1.67) measured on second EndoPAT on the contrast enhanced cardiac magnetic resonance parameters Total population (n=29) Endothelial dysfunction (RHI<1.67) p value d No (n=18) Yes (n=11) LVEdV (ml) a 138.6±26.9 139.2±26.5 137.5±28.9 0.87 LVEsV (ml) a 63.5±21.0 59.3±19.4 70.4±22.6 0.17 LVEF (%)a 53.9±8.4 56.6±8.1 49.5±7.2 0.025 Wall motion score index a 1.37±0.33 1.28±0.31 1.53±0.32 0.05 Oedema mass b 19.1 (19.0) 17.2 (14.1) 21.2 (28.6) 0.28 Transmural necrosis c 12 (38.7) 4 (22.2) 7 (63.6) 0.06 e Infarct mass Total b 11.6 (9.3) 10.1 (10.3) 17.5 (15.4) 0.08 Percent b 11.5 (13.7) 10.2 (7.6) 17.5 (21.8) 0.10 Indexed to APPROACH 3.2 (7.0) 2.7 (2.6) 4.9 (11.5) 0.10 Indexed to BARI 3.4 (5.8) 2.3 (2.7) 5.1 (11.5) 0.09 Salvage mass b 5.0 (14.0) 5.0 (8.8) 4.7 (27.6) 0.87 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Presented as number (%); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; e Yates correction In summary, the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT was related with larger infarcts, as assessed by troponin I release and ceCMR. These patients also had lower LVEF and worse wall motion score index and GLS in the acute echocardiogram and in the ceCMR, although these differences were not present in the follow-up exam echo. Results 148 6.2. Impact of the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT on microvascular reperfusion 6.2.1. Endothelial dysfunction and ST resolution on the ECG The residual total ST elevation and deviation and the percentage of resolution of these ST changes are presented in Table 68 , in Figure 25 and in Figure 26, according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation. Figure 25 – ST elevation and deviation resolution (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Although there was a tendency for higher resolution of ST elevation and deviation (particularly in post-PCI and 90 minutes ECGs) in patients without endothelial dysfunction, this difference was not statistically significant. Residual total ST elevation immediately after PCI and at 90 minutes was higher in patients with endothelial dysfunction and there was a trend for similarly worse results in residual ST deviation and residual ST elevation in these patients. The presence of Q waves was similar in both groups. 0 10 20 30 40 50 60 70 80 90 PostPCI 90 min 180 min PostPCI 90 min 180 min PostPCI 90 min 180 min Resolution of Total ST Elevation Resolution of Total ST Deviation Resolution of ST derivation with max elevation RHI > 1.67 RHI < 1.67 Results 149 Figure 26 – ST residual changes (median values) according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Table 68 – ECG ST resolution and residual changes according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Variable Total population (n=38) Endothelial dysfunction (RHI<1.67) p value d No (n=49) Yes (n=11) Immediately post-Angioplasty ECG a Total ST elevation b 3.0 (6.0) 2.5 (5.0) 5.0 (12.0) 0.048 % Resolution 74.5 (39.0) 76.5 (41.0) 60.5 (72.0) 0.23 Total ST deviation b 4.8 (6.0) 4.5 (5.0) 5.8 (14.0) 0.22 % Resolution 67.8 (39.0) 70.5 (39.0) 59.0 (78.0) 0.39 ST at derivation with max elevation) b 1.0 (2.0) 0.8 (2.0) 1.3 (3.0) 0.06 % Resolution 67.0 (33.0) 70.0 (40.0) 67.0 (52.0) 0.15 90 minutes post-Angioplasty ECG a Total ST elevation b 1.8 (4.0) 1.3 (4.0) 3.0 (7.0) 0.036 % Resolution 77.0 (29.0) 78.5 (28.0) 69.0 (53.0) 0.07 Total ST deviation b 2.0 (5.0) 1.8 (5.0) 3.0 (7.0) 0.191 % Resolution 78.0 (30.0) 81.0 (31.0) 71.0 (59.0) 0.181 ST at derivation with max elevation b 1.0 (2.0) 0.5 (1.0) 1.3 (2.0) 0.07 % Resolution 69.0 (31.0) 73.0 (37.0) 67.0 (44.0) 0.19 180 minutes post-Angioplasty ECG a Total ST elevation b 1.3 (4.0) 1.0 (3.0) 2.8 (5.0) 0.30 % Resolution 85.5 (26.0) 81.5 (22.0) 83.5 (52.0) 0.67 Total ST deviation b 1.5 (5.0) 1.3 (4.0) 2.8 (5.0) 0.74 % Resolution 84.5 (32.0) 83.0 (32.0) 85.5 (54.0) 0.80 ST at derivation with max elevation b 0.5 (1.0) 0.5 (1.0) 1.3 (2.0) 0.20 % Resolution 81.0 (30.0) 80.5 (29.0) 81.0 (40.0) 0.65 QS waves present 24 (63.2) 13 (59.1) 11 (68.8) 0.54 a Values expressed as median (interquartile range); b Values expressed in mV; c values expressed as n(%); d Mann-Whitney test for continuous variables with a non-normal distribution; e Chi-Square test 0 1 2 3 4 5 6 7 PostPCI 90 min 180 min PostPCI 90 min 180 min PostPCI 90 min 180 min Residual total ST elevation Residual total ST deviation Residual ST elevation max ST derivation RHI > 1.67 RHI < 1.67 p=0.048 p=0.036 Results 150 6.2.2. Endothelial dysfunction and angiographic indicators of microvascular reperfusion Corrected TIMI frame count and TIMI myocardial perfusion grade results according to the presence of endothelial dysfunction on the second EndoPAT evaluation are presented in Table 69. There was a trend toward higher values of cTFC and worse TMPG in patients with endothelial dysfunction, although it did not reach statistical significance. Table 69 – Angiographic indicators of microvascular reperfusion according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Variable Total population (n=38) Endothelial dysfunction (RHI<1.67) p value c No (n=22) Yes (n=16) Corrected TIMI frame count a 17.0 (7.0) 16.4 (7.0) 19.5 (12.0) 0.07 x TMPG 2-3 b 28 (73.7) 19 (86.4) 9 (56.3) 0.09 d a data presented as median (interquartile range); b data presented as n(%); c Mann-Whitney test; d Chi-Square test, with Yates correction. 6.2.3. Endothelial dysfunction and microvascular obstruction on the ceCMR Only 8 patients with the second EndoPAT evaluation had microvascular obstruction on the ceCMR: 2 (11.1%) in the group with RHI<1.67 and 6 (54.5%) in the group with RHI>1.67 (p=0.03). Likewise, the microvascular obstruction mass was higher in the group with RHI<1.67 (median value 5.3 vs. 6.8), although the difference was not significant (Table 70). Table 70 - Microvascular obstruction on the ceCMR according to the presence of endothelial dysfunction (RHI<1.67) on the second EndoPAT evaluation Total Population (n=29) Endothelial Dysfunction (RHI<1.67) p value c No (n=18) Yes (n=11) Microvascular obstruction MVO present a 8 (27.6) 2 (11.1) 6 (54.5) 0.03 d MVO mass b 6.1 (15.5) 5.3 (-) 6.8 (20.8) 0.29 a Presented as number (%); b Presented as median (interquartile range); c Mann-Whitney test for continuous variables with a nonnormal distribution, Chi-Square for categorical variables; d Yates correction In summary, there was a clear trend for worse angiographic and electrocardiographic indictors of microvascular reperfusion in patients with endothelial dysfunction on the second EndoPAT. The proportion of patients with MVO on the ceCMR was higher in patients with RHI<1.67 and there was a trend for higher MVO mass in these patients. Results 151 7. Secondary outcome – Extent of myocardial infarction and microvascular reperfusion according to IMR values In order to evaluate the impact of IMR on the extent of myocardial necrosis and its relation to invasive and non-invasive indicators of coronary microvascular reperfusion, patients were divided into two groups, according to the median value of IMR measured (24). To simplify the presentation of results and its discussion, patients with an IMR above the median 24 value will be classified as having coronary microvascular damage, as opposed to those with IMR values below 24. Where indicated, IMR as a continuous variable was also evaluated. In the following pages, the main patient characteristics according to IMR values will be shortly described, followed by the analysis of the impact of coronary microvascular damage. 7.1. IMR and baseline characteristics 7.1.1. IMR and patient’s main characteristics Patients with coronary microvascular dysfunction (IMR>24) were older (Table 71). In fact, there was a weak, but significant correlation between IMR and age (r=0.28, p=0.03). Table 71 – Main characteristics of patients according to median IMR Variable Total population (n=60) Coronary microvascular damage (IMR>24) p value c No (n=30) Yes (n=30) Physical characteristics Age (years) a 59.6±12.7 54.7±11.3 64.4±12.2 0.002 x Male gender b 48 (80.0) 26 (86.7) 22 (73.3) 0.20 x BMI a 27.5 ±4.0 27.4±4.4 27.6±3.7 0.79 x Waist circumference a 99.2±12.2 96.5±12.0 102.3±12.0 0.11 x Risk Factors and previous coronary disease b Hypertension 43 (71.7) 19 (63.3) 24 (80.0) 0.15 x Diabetes 15 (25.0) 7 (23.3) 8 (26.7) 0.77 x Dyslipidaemia 30 (50.0) 16 (53.3) 14 (46.7) 0.61 x Active smoking 26 (43.3) 16 (53.3) 10 (33.3) 0.12 x Previous angina 9 (15.0) 4 (13.3) 5 (16.7) 1.00 d Previous revascularization 2 (3.3) 1 (3.3) 1 (3.3) 0.49 d Previous medication b Aspirin 5 (8.3) 3 (10.0) 2 (6.7) 0.96 x ACEi/ARBs 23 (38.3) 12 (40.0) 11 (36.7) 0.87 x Beta-blockers 4 (6.7) 3 (10.0) 1 (3.33) 0.63 d Statins 9 (15.0) 3 (10.0) 6 (20.0) 0.44 d a Presented as mean±standard deviation; b Presented as number (%); c Independent t-test for continuous variables, Chi-Square for categorical variables; d Yates correction Results 152 There were no significant differences in gender, other physical characteristics or risk factors. NT-pro-BNP levels tended to be higher in patients with an IMR>24 and there was a weak but significant correlation between these 2 variables (r=0.33, p=0.015). Likewise, glucose and HbA1c levels on admission were higher in patients with IMR>24, and significant correlations were found between these laboratory tests and IMR (r=0.34, p=0.009 and r=0.67, p<0.001, respectively). Finally, there was a trend for higher values of hs-CRP in patients with coronary microvascular damage (Table 72). Table 72 – Laboratory result on admission according to median IMR Variable Total population (n=60) Coronary microvascular damage (IMR>24) p value c No (n=30) Yes (n=30) Creatinine (mg/dL) b 0.90 (0.31) 0.89 (0.17) 0.94 (0.44) 0.14 NT-pro-BNP (pg/mL) b 137.5 (255.5) 93 (165.5) 167.5 (222.8) 0.055 hs-CRP (mg(dL) b 0.44 (0.59) 0.33 (0.36) 0.49 (0.84) 0.09 Glucose (mg/dL) b 133.5 (61.0) 122.0 (48.8) 141.5 (60.0) 0.046 HbA1c (%)b 5.7 (1.1) 5.5 (0.6) 5.9 (3.5) 0.047 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution. 7.1.2. IMR and angiography/angioplasty variables There were no significant differences in ischemic (pain-to-balloon) or hospital-to-balloon times between patients with and without microvascular coronary damage (IMR>24). Main angiographic characteristics, including culprit artery, presence of multivessel disease and Syntax score were also similar in both groups (Table 73). Likewise, the area-at-risk, measured both by APPROACH and BARI scores, was similar. However, patients with an IMR>24 had a significantly higher prevalence of initial TIMI flow 01 (93.3% vs. 73.3%, p=0.038). Treatment options, including the use of mechanical aspiration and stent implantation technique were similar in patients with and without IMR>24 (Table 74). However, patients with lower IMR values were more likely to have received abciximab during the P-PCI procedure (36.7% vs 10.0%, p=0.015). Results 153 Table 73 – Ischemic times and angiographic characteristics according to median IMR Variable Total Population (n=60) Coronary microvascular damage (IMR>24) p valued No (n=30) Yes (n=30) Pain-to-balloon time (min)c 209 (148) 206 (107) 225 (163) 0.19 Door-to-balloon time (min) c 78 (45) 78 (34) 79 (59) 0.89 Culprit artery b Left anterior descending 28 (46.7) 14 (46.7) 14 (46.7) 0.94 Left circumflex 13 (21.7) 6 (20.6) 7 (23.3) Right coronary artery 19 (31.7) 10 (33.3) 9 (30.0) Multivessel disease b 25 (41.7) 15 (50.0) 20 (66.7) 0.19 Syntax score c 15.5 (10.0) 16.9 (10.0) 14.8 (8.0) 0.21 Area at risk scores a APPROACH score 27.8 (2.0) 27.8 (3.0) 27.8 (3.0) 0.48 BARI score 28.0±6.3 27.2±6.5 28.0±6.3 0.95 Initial TIMI flow 0-1 b 50 (83.3) 22 (73.3) 28 (93.3) 0.038 a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, Chi-Square for categorical variables; E Yates correction Table 74 – Angioplasty treatment options according to median IMR value Variable a Total Population (n=60) Coronary microvascular damage (IMR>24) p valueb No (n=30) Yes (n=30) Mechanical aspiration 26 (43.3) 14 (46.7) 12 (40.0) 0.60 Balloon pre-dilatation 35 (58.3) 15 (50.0) 20 (66.7) 0.19 Stent implantation 57 (95.0) 28 (93.3) 29 (96.7) 1.00 Balloon post-dilatation 23 (38.3) 12 (40.0) 11 (36.7) 0.79 Abciximab treatment 14 (23.3) 11 (36.7) 3 (10.0) 0.015 a Presented as number (%); b Chi-Square test 7.1.3. Multivariable analysis of IMR predictors Since there were several relevant differences in the populations with and without microvascular coronary damage and in order to clarify the independent predictors of IMR, multivariable analysis was performed, using IMR as a continuous variable. Variables statistically significant on the univariate analysis were included in the regression model: age, initial TIMI flow, admission glucose, admission HbA1c and abciximab treatment. Since glucose levels and HbA1c are interdependent variables, 2 different models were used, with these 2 blood tests. Results are shown in Table 75. Appendix – Papers 256 43. Bonetti PO, Pumper GM, Higano ST, Holmes DR, Kuvin JT, Lerman A. Noninvasive identification of patients with early coronary atherosclerosis by assessment of digital reactive hyperemia. J Am Coll Cardiol 2004;44(11):2137–41. 44. Flammer AJ, Anderson T, Celermajer DS, et al. The assessment of endothelial function: from research into clinical practice. Circulation 2012;126(6):753–67. 45. Damy T, Ratajczak P, Shah AM, et al. Increased neuronal nitric oxide synthase-derived NO production in the failing human heart. Lancet (London, England) 2004;363(9418):1365–7. 46. Marti CN, Gheorghiade M, Kalogeropoulos AP, Georgiopoulou V V, Quyyumi AA, Butler J. Endothelial dysfunction, arterial stiffness, and heart failure. J Am Coll Cardiol 2012;60(16):1455–69. 47. Winlaw DS, Keogh AM, Schyvens CG, Spratt PM, Macdonald PS, Smythe GA. Increased nitric oxide production in heart failure. Lancet 1994;344(8919):373–4. 48. Bauersachs J, Widder JD. Endothelial dysfunction in heart failure. Pharmacol Rep 60(1):119–26. 49. Dupuis J, Tardif J-C, Cernacek P, Théroux P. Cholesterol Reduction Rapidly Improves Endothelial Function After Acute Coronary Syndromes. Circulation 1999;99(25):3227 LP – 3233. 50. Gómez-Doblas JJ, Jiménez-Navarro MF, García-Pinilla JM, et al. [Effect of statin treatment begun early after acute myocardial infarction on endothelial function in patients with normal levels of cholesterol. VAATOPE Study (VAlue of ATOrvastatin in Postinfarction Endothelium)]. Med clinica 2006;126(9):325–8. 51. Akar JG, Al-Chekakie MO, Fugate T, et al. Endothelial dysfunction in heart failure identifies responders to cardiac resynchronization therapy. Heart Rhythm 2008;5(9):1229–35. 52. Enomoto K, Yamabe H, Toyama K, et al. Improvement effect on endothelial function in patients with congestive heart failure treated with cardiac resynchronization therapy. J Cardiol 2011;58(1):69–73. 53. Tesselaar E, Schiffer A, Widdershoven J, et al. Effect of cardiac resynchronization therapy on endotheliumdependent vasodilatation in the cutaneous microvasculature. Pacing Clin Electrophysiol 2012;35(4):377–84. Appendix – Papers 257 FIGURE LEGENDS Figure 1. Study flow chart Figure 2. Area under the curve of and peak values of troponin I (determined from 7 scheduled blood tests in the first 48 hours after P-PCI), according to the presence of endothelial dysfunction (RHI<1.67) Figure 3. Left ventricular ejection fraction (mean ± standard deviation) and wall motion score index (median and interquartile range) evaluated by Echo, according to the presence of endothelial dysfunction defined as an RHI<1.67 (*Independent T-test, **Mann-Whitney test) Appendix – Papers 258 TABLES Table 1. Baseline characteristics according to the presence of endothelial dysfunction (RHI<1.67). Variable Total Population (n=38) Endothelial Dysfunction (RHI<1.67) p value d No (n=22) Yes (n=16) Physical characteristics Age (years) a 60.0±13.7 60.1±12.4 59.8±15.7 0.94 x Male gender b 29 (76.3) 18 (81.8) 11 (68.8) 0.58 e BMI a 27.2±4.0 27.4±4.5 27.1±3.3 0.82 x Waist circumference a 100.0±11.5 101.1±12.6 98.5±9.9 0.55 x Risk Factors and previous coronary disease b Hypertension 26 (68.4) 16 (72.7) 10 (62.5) 0.50 x Diabetes 12 (31.6) 8 (36.4) 4 (25.0) 0.46 x Dyslipidemia 19 (50.0) 11 (50.0) 8 (50.0) 1.00 x Active smoking 13 (34.2) 9 (40.9) 4 (25.0) 0.31 x Previous angina 7 (18.4) 6 (27.3) 1 (6.3) 0.22 e Previous revascularization 2 (5.3) 1 (4.5) 1 (6.3) 0.61 e Previous medication b Aspirin / Clopidogrel 3 (7.9) 2 (9.1) 1 (6.3) 0.73 e ACEi/ARBs 13 (34.2) 10 (45.5) 3 (18.8) 0.11 x Beta-blockers 1 (2.6) 1 (4.5) 0 (0.0) 0.85 e Statins 4 (10.5) 2 (9.1) 2 (12.5) 0.89 e Admission laboratory values Creatinine (mg/dL) c 0.88 (0.32) 0.87 (0.54) 0.90 (0.19) 0.67 NT-pro-BNP (pg/mL)c 158.5 (305.0) 154 (365) 163 (250) 0.95 hs-CRP (mg(dL) c 0.57 (0.71) 0.29 (0.45) 0.81 (0.98) 0.06 Glucose (mg/dL) c 136.0 (63.0) 131.0 (60.0) 145.5 (81.8) 0.51 HbA1c (%)c 5.8 (1.3) 5.7 (1.4) 5.9 (2.1) 0.69 Ischemic times Pain-to-balloon time (min) c 209 (173) 209 (186) 211 (167) 0.94 Door-to-balloon time (min) c 75 (52) 79 (46) 57 (44) 0.06 Angiographic characteristics Culprit artery b Left anterior descending 21 (53.3) 11 (50.0) 10 (62.5) 0.89 e Left circumflex 7 (18.4) 5 (22.7) 2 (12.5) Right coronary artery 10 (26.3) 6 (27.3) 4 (25.0) Multivessel disease b 19 (50.0) 8 (36.4) 11 (68.8) 0.10 e SYNTAX score c 17.8±6.2 16.8±6.2 19.0±6.0 0.29 Initial TIMI flow 0-1 b 31 (81.6) 16 (72.7) 15 (93.8) 0.22 Area at risk scores c APPROACH score 27.8 (3.0) 28.1 (5.0) 29.7 (10.0) 0.08 BARI score 28.5 (6.0) 26.7 (7.0) 30.2 (10.0) 0.07 Angioplasty variables b Mechanical aspiration 16 (42.1) 9 (40.9) 7 (43.8) 0.86 Balloon pre-dilatation 21 (55.3) 13 (59.1) 8 (50.0) 0.58 Stent implantation 36 (94.7) 20 (90.9) 16 (100.0) 0.61 Balloon post-dilatation 24 (63.2) 14 (63.6) 10 (62.5) 0.94 Abciximab treatment 9 (23.7) 4 (18.2) 5 (31.5) 0.58 Mechanical aspiration 16 (42.1) 9 (40.9) 7 (43.8) 0.86 a Presented as mean±standard deviation; b Presented as number (%); c Presented as median (interquartile range); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, ChiSquare for categorical variables; e Yate’s correction Appendix – Papers 259 Table 2. Troponin I release according to the presence of endothelial dysfunction (RHI<1.67) Variable a, b Total Population (n=38) Endothelial Dysfunction (RHI<1.67) p value c No (n=22) Yes (n=16) Peak TnI values TnIpeak 95 (96) 67 (81) 118 (186) 0.024 TnIpeak (APPROACH) 25 (32) 17 (20) 34 (56) 0.009 TnIpeak (BARI) 24 (31) 17 (22) 33 (47) 0.008 TnIpeak (2 scores) 24 (31) 17 (21) 33 (55) 0.009 Area under the curve of TnI TnIAUC 1293 (1580) 1076 (1042) 2305 (2486) 0.012 TnIAUC (APPROACH) 403 (522) 315 (303) 664 (1080) 0.008 TnIAUC (BARI) 383 (448) 314 (326) 618 (799) 0.007 TnIAUC (2 scores) 393 (482) 314 (300) 641 (984) 0.007 a Presented as median (IQR); b Peak value and area under the curve (AUC) of 7 troponin I (TnI, in mg/dL) measurements performed in the first 48 hours after the primary angioplasty; total values and values indexed to the APPROACH, BARI or both are presented; c Mann-Whitney Test Appendix – Papers 260 Table 3. Echocardiographic results according to the presence of endothelial dysfunction (RHI<1.67) Variable Total Population Endothelial Dysfunction (RHI<1.67) p value c No Yes 2D measurements (n=26) (n=15) (n=11) LVTdV (ml) a 109.8±23.4 106.5±30.2 113.8±11.4 0.48 LVTsV (ml) a 55.9±12.6 51.1±12.3 61.7±10.9 0.047 LVEF (%)a 48.6±7.1 51.4±4.7 45.3±8.3 0.045 Wall motion score index b 1.44 (0.41) 1.35 (0.47) 1.77 (0.47) 0.006 Left atria volume (ml/m2) a 36.1±11.0 36.7±12.1 35.3±10.0 0.78 Doppler measurements (n=21) (n=11) (n=10) E/A ratio a 1.10±0.40 1.17±0.40 0.97±0.0.39 0.30 E/e’ ratio a 8.91±3.30 8.83±3.91 9.05±2.18 0.89 2D speckle tracking imaging (n=21) (n=11) (n=10) Global longitudinal strain a -13.16±2.35 -14.32±1.72 -11.89±2.35 0.014 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Independent t-test for continuous variables with a normal distribution. Mann-Whitney test for continuous variables with a non-normal distribution. LVEdV – left ventricular end diastolic volume; LVEsV – left ventricular end systolic volume; LVEF – left ventricular ejection fraction; WMSI – wall motion score index. Appendix – Papers 261 Table 4. ceCMR results according to the presence of endothelial dysfunction (RHI<1.67) Variable Total Population (n=29) Endothelial Dysfunction (RHI<1.67) p value d No (n=18) Yes (n=11) LVTdV (ml) a 138.6±26.9 139.2±26.5 137.5±28.9 0.87 LVTsV (ml) a 63.5±21.0 59.3±19.4 70.4±22.6 0.17 LVEF (%)a 53.9±8.4 56.6±8.1 49.5±7.2 0.025 Wall motion score index a 1.37±0.33 1.28±0.31 1.53±0.32 0.05 Edema mass b 19.1 (19.0) 17.2 (14.1) 21.2 (28.6) 0.28 Transmural necrosis c 12 (38.7) 4 (22.2) 7 (63.6) 0.06 e Infarct mass Total b 11.6 (9.3) 10.1 (10.3) 17.5 (15.4) 0.08 Percent b 11.5 (13.7) 10.2 (7.6) 17.5 (21.8) 0.10 Indexed to APPROACH 3.2 (7.0) 2.7 (2.6) 4.9 (11.5) 0.10 Indexed to BARI 3.4 (5.8) 2.3 (2.7) 5.1 (11.5) 0.09 Salvage mass b 5.0 (14.0) 5.0 (8.8) 4.7 (27.6) 0.87 Microvascular obstruction MVO present c 8 (27.6) 2 (11.1) 6 (54.5) 0.03 e MVO mass b 6.1 (15.5) 5.3 (-) 6.8 (20.8) 0.29 a Presented as mean±standard deviation; b Presented as median (interquartile range); c Presented as number (%); d Independent t-test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, ChiSquare for categorical variables; e Yate’s correction Appendix – Papers 262 FIGURES Figure 1 Appendix – Papers 263 Figure 2 Figure 3