Association of aortic stiffness early post myocardial infarction with left ventricular remodelling

Adverse left ventricular (LV) remodelling after myocardial infarction is associated with heart failure. We investigated whether aortic stiffness during acute ST‐segment elevation myocardial infarction is associated with LV remodelling at long‐term follow‐up.


| INTRODUCTION
It is well known that increased aortic stiffness is an independent risk factor for cardiovascular events. 1,2From a pathophysiological point of view, this is explained by the fact that an increase in aortic stiffness results in elevated central pulse pressure, left ventricular (LV) afterload and decreased coronary artery perfusion. 3Among the different methods to estimate aortic stiffness, carotid to femoral pulse wave velocity (PWV) is considered as the gold standard method due to its ease of identification and its reliability. 4Myocardial ischemia and increased LV afterload lead the heart to adverse remodelling, which is associated with the onset of heart failure. 5Interestingly, after acute myocardial infarction, a non-contractile and expanding infarcted area of scar tissue is formed.This expansion causes an increased volume load, which subsequently augments the pressure load exerted on non-infarcted regions leading to adverse LV remodelling.Long-term progressive LV remodelling after myocardial infarction can last up to 2 years after the index event and is linked to increased cardiovascular mortality. 6Therefore, the evaluation of LV remodelling parameters has significant therapeutic and prognostic implications.
However, it is not clear whether the impairment of ventricular-arterial interaction at the early phase of STsegment elevation myocardial infarction (STEMI) because of increased aortic stiffness and/or impaired LV deformation is linked to adverse LV remodelling.Based on the aforementioned observations, in the present study, we hypothesized that aortic stiffness in addition to LV myocardial deformation were associated with long-term LV remodelling in patients with STEMI.Thus, in a prospective study, we measured carotid to femoral pulse wave velocity, LV global longitudinal strain and their ratio as a surrogate index of ventriculararterial interaction, within 48 h of STEMI post-primary percutaneous coronary intervention and after 2 years of the index event, aiming to examine their predictive value for LV remodelling during a 2-year follow-up period.

| Study design and population
The present study was a prospective observational singlecentre study conducted at the Second Department of Cardiology in Attikon University Hospital (Athens, Greece).Patients of this study were previously included as a control group in a randomized study which investigated the acute and chronic effects of remote ischemic conditioning on aortic stiffness and LV remodelling in two groups of patients after acute myocardial infarction who underwent two different remote conditioning protocols in comparison to a control group of patients without remote conditioning intervention which is now included in the current study. 7n the current study, 162 consecutive hospitalized patients with acute myocardial infarction were assessed for eligibility by the attending cardiologists (I.I. and K.K.) in the period between 2015 and 2019.Inclusion criteria were male or female patients aged 35 to 85 years with STEMI diagnosed within 48 h and after successful primary percutaneous coronary intervention (PCI).The diagnosis of STEMI made on the basis of typical electrocardiography changes and/or chest pain that was associated with elevated cardiac biomarkers. 8Exclusion criteria included history of previous known coronary artery disease, previous PCI or coronary artery bypass grafting, Killip class >2 or cardiogenic shock during the index event, treatment with nitrates, arrhythmias, including atrial fibrillation, moderate or severe valve disease, as well as chronic inflammatory diseases and malignancies.Out of the 162 patients, 27 were excluded from the current study due to previous PCI (n = 15) or coronary artery bypass grafting (n = 6), Killip class >2 (n = 4) and because of chronic inflammatory disease (n = 2).Hence, 135 patients underwent estimation of arterial and cardiac function on the same day within 48 h after STEMI and primary PCI (mean time: 36 ± 12 h).Of those, 10 patients were excluded because of poor image quality for strain analysis.Thus, 125 patients were scheduled for re-examination at 2 years after index events.Two years after the index hospitalization, 5 patients died and 11 patients were lost at follow-up.Finally, 109 patients were included in the analysis (Figure 1).The mean time from ischemic event to re-examination was 24 ± 1 months.The decision to re-evaluate patients after 2 years was based on the following reasons: (a) in our previous study, we observed that left ventricular end-systolic volume had a statistically non-significant reduction in revascularized post myocardial infarction patients under standard treatment without remote conditioning intervention at 1 year after acute myocardial infarction 7 and (b) studies report that LV remodelling may last up to 2 years post infarction. 6he study was conducted according to the principles of the Declaration of Helsinki, and protocol was approved by the hospital's Ethics Committee (Approval number: 301/27-6-2014).Additionally, all patients gave written informed consent prior to participation in the study.

| Aortic stiffness assessment
Aortic stiffness was estimated by carotid to femoral pulse wave velocity (PWV, m/s) using tonometry (Complior SP, Alam Medical).Two noninvasive sensors were used to capture carotid and femoral waveforms.The physical distance between the two arterial sites was assessed by a tape measure.Pulse wave velocity was calculated as the direct distance from the carotid to femoral site divided by the transit time between two waveforms.Normal value of PWV is reported to be <10 m/s for indirect method (common carotid artery -common femoral artery × 0.8). 9any different methods for the evaluation of aortic PWV have been proposed.However, a recent invasive validation study of several devices demonstrated that only methods estimating carotid to femoral PWV (namely Complior and SphygmoCor) presented a strong correlation and agreement with invasively measured aortic PWV, and thus, they seem to be reliable approaches for assessment of aortic stiffness and may be considered interchangeable. 10

| Echocardiography
All patients were studied by transthoracic echocardiography using a Vivid E95 (General Electric Medical Systems) ultrasound system.Studies were digitally stored in a computerized station (EchoPac General Electric 203) and were analysed by two observers (D.T. and G.P.) who had no access to the patients' clinical and laboratory data.Inter-and intra-observer variabilities were calculated as a coefficient of variation (mean standard deviation/mean × 100).In order to evaluate inter-observer variability, data from the first 20 patients were analysed by the two investigators.
Intra-observer variability between the first and second measurements was determined by the same investigator.

| Two-dimensional strain analysis
Two-dimensional strain was assessed by the speckletracking imaging method (EchoPac PC 203, General Electric Healthcare, Horten, Norway).The LV global longitudinal strain (GLS, %) was measured using a 17-segment division of the left ventricle from the apical four-, two-and three-chamber views, as previously published. 11The normal GLS is considered to be −22.5 ± 2.7%. 12The inter-and intra-observer variability of GLS was 7% and 10%, respectively.Left atrial (LA) function was assessed using speckletracking strain imaging in apical four-and two-chamber views.Reservoir LA strain was calculated as peak longitudinal strain during ventricular systole according to previously published methodology. 13The normal reservoir LA strain is reported to be 39.4% (95% CI: 38%-40.8%). 14

| Ventricular-arterial interaction
The ratio of carotid to femoral PWV to LV GLS (PWV/GLS, m/s%) was calculated as a surrogate index of ventriculararterial interaction as previously published. 9The ratio has negative values due to negative GLS values, and the normal values are considered to be −.42 ± .12m/s%. 15

| Left ventricular remodelling
Left ventricular remodelling was determined by twodimensional echocardiography using a Vivid E95 (General Electric Medical Systems, Horten, Norway) ultrasound system.Left ventricular end-diastolic (LVEDV) and endsystolic (LVESV) volumes were measured utilizing the modified biplane Simpson's technique form the apical four-and two-chamber views within 48 h post PCI and after 2 years. 7Adverse LV remodelling can be defined as an increase in LVESV >15% at follow-up compared with baseline values. 16A cutoff of >15% decrease from the baseline in LVESV was considered as a criterion of reverse LV remodelling, as this value is a validated reverse remodelling index in patients with ischemic heart disease. 17,18

| Statistical analysis
Statistical analysis was performed using the Statistical Package for Social Sciences (IBM SPSS Statistics for Windows, Version 26.0.Armonk, NY, USA) and STATA (Stata Statistical Software: Release 16.College Station, TX, USA).Continuous variables are expressed as mean ± standard deviation or as median with interquartile range (first quartile-third quartile) according to their distribution.Differences in continuous variables were tested by the Student's t-test and Mann-Whitney U-test, as appropriate.Categorical variables are expressed as numbers with corresponding percentages and are compared by the chisquare test.Simple correlations between the continuous variables were evaluated using parametric (Pearson's r) or non-parametric (Spearman's rho) correlation coefficients.Univariate and multivariable linear regression analyses were performed to investigate the associations of clinical, vascular and echocardiographic variables with LV volumes change between baseline and 2 years of followup and the corresponding unstandardized coefficient beta and 95% confidence intervals (CIs) were calculated.Univariate and multivariate logistic regression analyses were performed to investigate the associations of clinical, vascular and echocardiographic variables with LV remodelling (>15% decrease from the baseline in LVESV at follow-up), and their corresponding odds ratio and 95% confidence intervals were calculated.All variables with a p-value < .05 in univariate analysis were included in the multivariable model.The C-statistic was calculated to evaluate the predictive value of the examined variables for the presence or absence of reserve LV remodelling.All statistical tests were two-tailed, and a p-value < .05 was considered significant.

| RESULTS
Demographic, clinical and laboratory data of the whole study population are presented in Table 1.One hundred nine patients (88 men and 21 women) with a mean age ± standard deviation of 58 ± 10 years were included in the present study within 48 h of STEMI post-primary PCI.Out of the 109 patients, 52 (48%) were diagnosed with an anterior STEMI, 46 (42%) had single-vessel, 50 (46%) double-vessel and 13 (12%) triple-vessel coronary artery disease.The median time from onset of symptoms to first balloon inflation during PCI was 180 min [interquartile range (IQR): 133-279], and median time from first medical contact-to-balloon inflation was 104 min [IQR: 80-131].Median high-sensitivity cardiac troponin T (hs-cTnT) was 3950 ng/mL [IQR: 982-7195].All patients received the appropriate treatment according to current guidelines throughout the 2-year follow-up period. 19More specifically, the treatment of STEMI patients included antiplatelet (acetylsalicylic acid and clopidogrel) and anticoagulant therapy (enoxaparin).Additional intravenous bolus infusions of unfractionated heparin were administered during PCI.Furthermore, all patients were treated with angiotensin-converting enzyme inhibitors, βblockers and statins.When the total study population was divided into two groups according to the presence or not of reverse LV remodelling (>15% decrease from the baseline in LVESV), there were no significant differences in the baseline demographic and clinical characteristic between the two groups, except for baseline hs-cTnT which was higher in patients without reverse LV remodelling compared with those who presented reverse LV remodelling at 2-year follow-up (p = .042;Table 1).
On the contrary, in patients with LVESV reduction <15%, LV ejection fraction showed a non-significant reduction (44 ± 9 vs. 42 ± 10%, p = .123)at 2 years after acute myocardial infarction (Table 2).By stepwise linear regression analysis using both for- GLS ratio and LA strain were significantly associated with reverse LV remodelling (Table 3).Multivariable analysis revealed that baseline PWV (odds ratio: .735,95% CI: .381-.967, p = .046),GLS (odds ratio: .846,95% CI: .606-.994, p = .048)and PWV/GLS ratio (odds ratio: 2.733, 95% CI: 1.748-3.629,p = .034)were independently associated with LV remodelling after STEMI.The multivariable model included age, sex, anterior myocardial infarction, baseline values of hs-cTnT, baseline PWV, GLS, PWV/GLS, LA strain, LV volumes, LV ejection fraction, systolic and diastolic blood pressure, heart rate and concomitant medications and utilized both forward and backward procedure to assess the variables with independent association with LV remodelling.Moreover, addition of baseline PWV, GLS and PWV/GLS separately in a multivariable model including age, sex, anterior myocardial infarction, baseline hs-cTnT, LV volumes, LV ejection fraction, systolic and diastolic blood pressure, heart rate and concomitant medications showed that lower baseline PWV and a less impaired baseline GLS and PWV/GLS were predictive of reverse LV remodelling at 2 years with a C-statistic of .748,(95% CI: .597-.903) for PWV, .711,(95% CI: .519-.902) for GLS and .787(95% CI: .678-.892) for PWV/GLS ratio.In the present study, we have demonstrated that all patients with revascularized STEMI had improved myocardial and vascular function at 2 years after the index event, which may be explained by the myocardial recovery after timely and successful revascularization for the improved cardiac and by the response to treatment for vascular function over the follow-up period.Patients with LVESV reduction >15%, (a criterion of reverse LV remodelling), had lower baseline pulse wave velocity and higher baseline LA strain and better LV myocardial strain than those without significant LVSEV reduction reflecting a better ventricular-arterial interaction early after acute myocardial infarction.Additionally, 2 years after index event, patients with LVESV reduction >15% showed further improvement of PWV, GLS, PWV/GLS ratio and reservoir LA strain compared with patients with LVESV reduction <15% suggesting a continuing improvement of ventricular-arterial interaction throughout the 2 years of follow-up.This finding implies that apart from aortic stiffness early post myocardial infarction, improvement of aortic elastic properties during followup may have also contributed to reverse LV remodelling.Interestingly, baseline values of hs-cTnT, PWV, GLS, PWV/GLS ratio and LA strain had an important impact in reverse LV remodelling.However, multivariable analysis revealed that baseline PWV, GLS and PWV/GLS ratio are superior to other parameters for the prediction of LV remodelling after STEMI.
Left ventricular remodelling is characterized by a sum of genetic, molecular, cellular and interstitial changes with subsequent alterations in cardiac size, shape and function after myocardial injury such as an ischemic event. 20A recent study demonstrated that the development of adverse LV remodelling post myocardial infarction is related with an 8.5-fold increase in the risk of cardiovascular events during a 48-week follow-up period. 21Conversely, even minor reductions in LV remodelling can be associated with a decreased risk of heart failure and cardiovascular death. 22On the contrary, reverse LV remodelling is defined as a process characterized by a decrease in LV volumes resulting in a significant restoration of systolic and diastolic cardiac function after timely revascularization. 23 T A B L E 3 Univariate and multivariable logistic regression analysis to predict reverse LV remodelling.
5][26] Indeed, in our study, hs-cTnT was higher in the group with LVESV reduction <15% at 2 years follow-up and univariate analysis revealed that it was associated with reverse LV remodelling.Additionally, aortic PWV, as assessed by magnetic resonance imaging within 1 week after acute revascularized STEMI, independently predicted infarct size reduction as evaluated also by cardiac magnetic resonance at 4 months post myocardial infarction. 27urthermore, Oleynikov et al. have showed the predictive value of LV-arterial coupling, defined as the ratio of arterial elastance to LV end-systolic elastance, in the development of adverse LV remodelling (relative risk = 1.96, 95% CI: 1.11-3.46,p = .020). 21However, a recently published study demonstrated that PWV/GLS ratio was importantly associated with cardiovascular factors regardless of age and thus should preferably be used to assess ventricular-arterial interaction over other traditional echocardiography-derived markers. 15In line with these studies, we found in the current study that higher baseline PWV, and impaired PWV/ GLS ratio, as a surrogate marker of ventricular-arterial interaction, are associated with absence of reverse LV remodelling at 2 years post myocardial infarction.Several factors may affect aortic stiffness.Multiple studies have showed an age-related increase in PWV while the relationship between sex, race and aortic stiffness is more complex and varies with age.Additionally, comorbidities including hypertension, hyperlipidaemia, obesity, metabolic syndrome, diabetes mellitus, as well as lifestyle risk factors, such as smoking, alcohol, diet habits and limited physical activity, have a detrimental effect on aortic stiffness.Furthermore, accumulating evidence shows that several biomarkers including natriuretic peptides, biomarkers involved in the inflammatory cascade and the renin-angiotensinaldosterone system, may be associated with increased aortic stiffness. 28On the contrary, calcium channel blockers, angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers have been shown to have a beneficial effect on arterial stiffness in hypertensive subjects. 29In recent years, novel antidiabetic agents, namely glucagon-like peptide-1 receptor agonists and sodium-glucose cotransporter-2 inhibitors, 11 as well as monoclonal antibodies have targeted biologic pathways involved in autoimmune diseases, seem to improve aortic stiffness leading to reduced cardiovascular risk. 30It has been demonstrated that elevated aortic stiffness, as assessed by PWV, increases LV afterload and decreases coronary perfusion pressure during diastole with resultant the development of ischemia in sub-endocardial layers and ventricular-arterial decoupling even in patients with non-obstructive coronary artery disease. 9,31In addition, increased aortic stiffness affects LV diastolic filling pressures contributing to LA dysfunction as estimated by the sensitive echocardiographic marker of LA strain. 13Experimental evidence shows that even transient increases in LV afterload may detrimentally affect LV remodelling after myocardial infarction. 32Previous studies have revealed the importance of LV longitudinal strain as a strong independent predictor of LV remodelling. 25,33However, these studies differed in infarct location and duration of follow-up.Further expanding these findings, the present study is the first in our knowledge to show the association of reverse LV remodelling at 2 years post-STEMI with the presence of a better ventricular-arterial interaction, as assessed by PWV, GLS and PWV/GLS ratio within 48 h post-primary PCI in STEMI.Reduced aortic stiffness post myocardial infarction indicates an important reduction of LV afterload, which in turn may have beneficial effect on myocardial perfusion contributing to the improvement of LV myocardial deformation and consequently to improved ventricular-arterial interaction, systolic performance and LV diastolic function, as estimated by improved LA strain in our study. 9,34,35Thus, it seems that aortic stiffness plays a key role in ventricular-arterial interaction post myocardial infarction in our study and by multivariable analysis a low baseline value of PWV was an independent predictor of post-infarction reverse LV remodelling at 2 years of follow-up.
In our study assessment of aortic stiffness, LV myocardial deformation and ventricular-arterial interaction as estimated by PWV, GLS and PWV/GLS early after the onset of STEMI conferred significant incremental value for the subsequent LV remodelling compared to other classic clinical and conventional echocardiographic parameters such as LV volumes and ejection fraction.Thus, early determination of aortic stiffening and ventricular-arterial interaction may provide important prognostic information for the progression to heart failure post-STEMI, and may become a therapeutic target to halt adverse LV remodelling and progression to heart failure by utilizing a more aggressive treatment at the early phase post myocardial infarction.

| Limitations
There are some limitations in the present study.First, our study had a moderate sample size, so our results need to be confirmed by larger data sets.Second, the relatively increased percentage of male patients in the current study should be acknowledged as a limitation as female patients may be underrepresented.Large scale, multicentre and prospective studies are necessary to further explore the association of arterial and cardiac function with longterm LV remodelling in patients after acute myocardial infarction.

| CONCLUSIONS
In patients with revascularized STEMI, aortic stiffness, myocardial deformation and ventricular-arterial interaction with 48 h of the acute event are independently associated with the extent of LV remodelling 2 years after the ischemic event.Thus, early assessment of aortic stiffness, LV myocardial deformation and their ratio in the acute phase of myocardial infarction may provide valuable information for patient risk stratification and may provide therapeutic guidance to prevent progression to heart failure.

F I G U R E 1
Flow diagram of the study cohort.PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting.

F I G U R E 2
The percentage reduction from baseline of left ventricular end-systolic volume (ΔLVESV) with a cutoff of 15% in relation to the percentage changes from baseline (Δ) in (A) pulse wave velocity (PWV), (B) global longitudinal strain (GLS), (C) PWV/GLS ratio and (D) left atrial (LA) strain at 2-year follow-up.Data shown are means ± standard deviation.
T A B L E 1Note: Data are presented as number (%), means ± standard deviation, or median values [first quartile -third quartile].Abbreviations: ΔLVESV, percentage reduction from baseline of left ventricular end-diastolic volume; BMI, body mass index; CAD, coronary artery disease; CRP, C-reactive protein; hs-cTnT, high-sensitivity cardiac troponin T; LAD, left anterior descending artery; LCX, left circumflex artery; RCA, right coronary artery; WBC, white blood cell count.T A B L E 2 Progression of aortic stiffness and echocardiographic markers of myocardial function within 2 years of the index event.
Studies Concomitant medications included antiplatelets, angiotensin-converting enzyme inhibitors, βblockers and statins.The multivariable model included age, sex, anterior myocardial infarction, baseline values of hs-cTnT, baseline PWV, GLS, PWV/GLS, LA strain, LV volumes, LV ejection fraction, SBP, DBP, heart rate and concomitant medications and utilized both forward and backward procedure to assess the variables with independent association with LV remodelling. Note: