Endovascular aortic repair impact on myocardial contractility: A prospective study

This study aimed to estimate if the altered sphygmic wave transmission may affect the left ventricular (LV) contractile function in patients undergoing endovascular aortic repair (EVAR).


| INTRODUCTION
Even if the endovascular aortic repair (EVAR) has become the predominant treatment for abdominal aortic aneurysms (AAA) over the past few years, [1][2][3] there are still concerns about the long-term results of EVAR and several studies have reported that it can shorten long-term survival, also due to cardiac complications, compared to open repair. 4,5 Although the underlying mechanisms still need to be cleared, the haemodynamic consequence of EVAR might contribute to the risk of cardiac events. Indeed, the stiffness of the stent-graft is known to modify the aortic compliance and alters both forward transmission and backward reflection of the sphygmic wave. [6][7][8][9][10][11] This interference could potentially affect the systolic afterload to left ventricular (LV) contraction and the coronary driving pressure pattern, and altering myocardial perfusion response to physical or emotional stresses, especially in patients with co-existing coronary artery disease (CAD). [12][13][14] From the methodological point of view, the altered transmission of the pressure wave hampers the accuracy of the clinical evaluation of LV function by the ejection fraction (EF), since the uncertain pressure at the aortic valve (AV) does not permit to correctly account for the vascular resistance to systolic output.
This potential limitation has been at least partially solved by Suga and Sagawa with the introduction of the ratio between LV end systolic pressure (ESP) and volume (ESV) also termed maximal systolic myocardial stiffness (MMS), that so far represents the gold standard descriptor of LV contractility. 15 Despite its proven accuracy, the clinical application of MMS has been profoundly hampered by the technical complexity of the simultaneous monitoring of LV volume and pressure throughout the cardiac cycle and thus the need of cardiac catheterisation. For this reason, the use of MMS has been so far limited to pathophysiological studies aiming to provide precise estimations of LV contractility.
ECG-gated acquisition of myocardial perfusion imaging (MPI) can overcome many of the limitations affecting the echocardiographic estimation of LV volume at welldefined phases of the cardiac cycle. 17,18 This nuclear imaging modality, indeed, permits to directly estimate LV volumes at precise times after the R wave and thus to accurately match the ESP, estimated at the AV, and the corresponding ESV.
In the present study, we applied this approach to estimate whether the altered transmission of the sphygmic wave affects the LV contractile function in a series of consecutive patients that underwent EVAR for AAA.

| METHODS
We performed a single-centre prospective study on patients attending EVAR for AAA at the Vascular and Endovascular Surgery Clinic -San Martino Hospital, Genoa (Italy). The present study was approved by the Regional Ethics Committee (Comitato Etico Regione Liguria -protocol number 29/12) and signed informed consent was obtained by each patient included in the study. Reporting of the study conforms to broad EQUATOR guidelines. 19 Each patient underwent preoperative cardiac evaluation encompassing a stress-rest MPI with gated single photon emission computed tomography (SPECT) associated with arterial stiffness measurement. MPI was performed both before and 6 months after EVAR. In the case of a positive preoperative SPECT, the cardiologist would have indicated a following coronary angiogram and an eventual further revascularisation, thus excluding the patient from the study. According to the American College of Cardiology/American Heart Association (ACC/AHA) guidelines, five factors were considered in determining whether a patient should undergo further cardiac testing for risk stratification. 20 These factors are (1) recent coronary revascularisation, if any; (2) distance from the last favourable cardiac evaluation; (3) presence of patient comorbidities, classified as major, intermediate, or minor clinical predictors; (4) functional status; and (5) the risk of the proposed surgery. This stage will be entrusted to a cardiologist dedicated to the evaluation of patients.

| Inclusion and exclusion criteria
The inclusion criteria were the indication to elective EVAR for an asymptomatic AAA in the absence of previous aortic interventions. Indication to EVAR was made taking into consideration the most recent guidelines and specialists' choice. Exclusion criteria were symptomatic or ruptured AAA, non-atherosclerotic aneurysm, paroxysmal or chronic atrial fibrillation, and patients undergoing coronary revascularisation between the two SPECT.

| SPECT protocol
Stress-rest MPI was performed according to a one-day protocol. For stress scan, a bolus of 200 MBq of 99mTcsestamibi was administered intravenously 2 min after dipyridamole infusion (.56 mg/kg body weight over 4 min). Rest imaging was performed hours later and implied the bolus administration of 550-600 MBq of the same tracer.
Imaging was acquired 30-45 min after tracer injection using a dual-head gamma-camera (Discovery NM630, GE Healthcare) equipped with parallel hole low-energy highresolution collimators with an energy window (10%) centred over 140 keV. Heads were set at L position and a 180° arc was covered according to a step and shoot modality; each view lasted 45 and 25 s for stress and rest imaging, respectively. ECG-gating was set to acquire 16 frames per cardiac cycle. R-R intervals were accepted only if the duration ranged as mean ± 15%; in all cases, the first postectopic beat was excluded to avoid the interference by post-extrasystolic potentiation.
Obtained images were processed using the commercially available validated QPS/QGS software to obtain the list of the average 16 LV volumes, and their time frame during the average cardiac cycle, throughout the whole acquisition. LV ejection fraction (LVEF) was computed according to the standard method, that is,: LVEF = (EDV − ESV) / EDV with EDV = first (end-diastolic) and ESV = minimal (end-systolic) volumes.
To better define cardiac contractility, the curve of LV volumes was defined and matched with the corresponding data of the central aortic pressure curve. Only images acquired under resting conditions were analysed with this approach. Indeed, with the 1-day protocol, this condition is associated with a significant improvement in counting statistics that allows an accurate definition of myocardial borders in all investigated frames. The availability of pressure and LV volume data during the systolic opening of AV permitted us to estimate the MMS index of ventricular contractility, which was evaluated by the ratio of systolic central aortic pressure and LV volume in the frame in which this latter variable was at its minimum value.

| Central aortic waveform measurement
The pressure wave curve in the central aorta was estimated using the SphygmoCor XCEL device (AtCor Medical). 21 This tool uses a standard brachial cuff to measure brachial systolic and diastolic pressures, and to capture the pressure waveform in the brachial artery. The curve was then analysed by the device to derive the central aortic waveform. 16 Besides the conventional analysis of pressure reported in Figure 1, this approach permits to define the pressure waveform in the ascending aorta with each time interval lasting less than 10 ms. The parameters evaluated with the SphygmoCor XCEL device (AtCor Medical) were SP (central aortic systolic pressure), AP (central augmented pressure), PP (central aortic pulse pressure), AIx (augmentation index), AIx75 (normalised augmentation index), MAP (mean arterial pressure), RWTT (reflected wave transit time) and RTI (reflection time index). In detail, these parameters are described as follows: SP, that is, the maximum pressure during aortic ejection; AP, that is, the difference between two pressure peaks during systole (P2-P1); PP, that is, the height of the aortic pressure waveform -PP > 50 mmHg has been found to be predictive of cardiovascular diseases; MAP, that is, the average aortic pressure in a pulse. The AIx was derived to evaluate the physical state of the arterial system and it was calculated as the ratio of AP to PP, multiplied by 100 (AI = (AP/PP) × 100). The HR was measured as well and AIx was normalised for 75 beats per minute HR (AIx75) to compare patients with different HR and the stiffness of the same patient. The RWTT, which is the time from the beginning of the systole to the arrival of the reflected wave, was computed as a marker of large artery stiffness. The RTI was obtained by normalising the RWTT to the duration of the cardiac cycle.
Arterial pressure waveform was acquired in the right brachial artery every 2 min throughout the entire MPI duration both at stress and at rest. The obtained data were averaged and segmented into the same 16 intervals to match with the corresponding time frames of calculated LV volumes measured. ESP was thus set to the time of minimal ESV, and the MMS was then calculated. F I G U R E 1 Example of aortic pressure waveform (red), with incident wave (blue) and reflected wave (green) highlighted. AP, augmented pressure (P2-P1); PP, pulse pressure; RWTT, reflected wave transit time; SP, systolic pressure.

| EVAR technique and follow-up after EVAR
According to the internal protocol and to the most recent European guideline, patients that undergo EVAR are subjected to a lifelong surveillance program. 3 A thoracoabdominal computed tomography angiography is performed within 30 days after EVAR and, in the absence of any complication, we perform an annual follow-up thereafter. Whenever an inadequate seal, a sac expansion or an endoleak occurs at the first postoperative control, it should be evaluated for reintervention or followed with at least an annual Doppler ultrasound.

| Statistical analysis
Data are shown as counts, percentages, median with a range and mean ± SD, where appropriate. Statistical differences were analysed by means of Student t test for paired samples or non-parametric Mann-Whitney test. p-values < .05 were considered statistically significant in all statistical tests. All the analyses were performed using Matlab software 2019b -Statistics and Machine Learning Toolbox (Mathworks Inc.).

| RESULTS
Between 2018 and 2020, 24 patients matched the inclusion and exclusion criteria. After preoperative examination, 9 of 24 patients were subsequently excluded: one patient refused the treatment, four patients underwent fenestrated or branched endovascular aortic repair and four patients showed reversible perfusion defects at MPI and were thus submitted to coronary revascularisation before EVAR.
Mean age of included patients was 76.6 ± 6.3 years and all of them were male. Demographics data and perioperative medication management were respectively reported into Tables 1 and 2. Beta blockers were the most prescribed anti-hypertensive medications before the procedure (5, 33.3%) and nine patients (60%) were already under statin therapy. Most patients were already under single antiplatelet therapy before the intervention (9, 60%) while six patients (40%) started a dual antiplatelet therapy at the discharge.
The reflection time index (RTI), which is the RWTT normalised to the cardiac period, also slightly decreased from pre-operative to 6-month post-operative examination both at stress (.13 ± .02-.12 ± .03) and at rest (.13 ± .02-.11 ± .02) acquisitions. No significant changes were observed in SP, PP, AP, AIx, AIx75, and HR parameters between preoperative and post-operative investigations (Table 3). Complete stress and rest myocardial perfusion data were available in all 15 patients both before and after EVAR ( Figure 2C). At rest images, ESV increased during the 6 months follow-up from 34 ± 9 mL to 39 ± 8 mL (p = .02); by contrast, the increase in EDV did not reach the statistical significance (from 85 ± 34 mL to 89 ± 29 mL, respectively, p = .6). LVEF was not apparently modified since it was 65 ± 6% and 67 ± 6% before and after treatment, respectively (p = ns). Nevertheless, the MMS significantly decreased from 3.6 ± 1.5 to 2.66 ± .74 mmHg/mL (p = .03).

| DISCUSSION
The main finding of the present study is that the altered transmission of the sphygmic wave induced by EVAR is associated with an early contractile impairment.
Since this alteration is obviously transmitted throughout the entire arterial tree, this impairment might affect the pressure in the proximity of the AV and thus the afterload to LV contraction. This potential limitation was confirmed by the evidence that the segmental increase in aortic stiffness, caused by EVAR, is followed by the appearance of both LV hypertrophy, and left atrial enlargement, that occur early after surgery and in the absence of any change of blood pressure. 6 According to Marketou et al., 14 AAA repair is related to both an increase in aortic stiffness and a reduced cardiac systolic function. Arterial stiffness evaluated using pulse wave velocity has been suggested to be one of the key factors for determining sac behaviour after EVAR. 22 Moreover, experimental studies have also suggested that elevated pulse wave velocity after endovascular treatment results in decreased aortic compliance and increased aortic impedance in animal models. 23 In our experience, both these alterations might indicate an impairment in the contractile function of the LV myocardium despite the absence of any decline in LVEF after aortic repair, at least in the short term. In agreement with this clinical notion, LVEF remained unchanged for the 6 months follow-up period considered in our study. 14 Despite the invariant LVEF value, the post-surgical contractile impairment was indicated by the decrease in MMS. In its original formulation, this index is calculated by the complete definition of the LV pressure-volume loop and thus requires the simultaneous monitoring of these variables within the LV chamber and thus cardiac catheterisation. 24,25 Our approach surrogates this complete definition by limiting the analysis to the systolic phase, which allows us to consider the absence of any pressure gradient across the AV. This analysis implied an accurate definition of the LV volume curve during the cardiac cycle to be matched with the corresponding trend of central aortic pressure. It obviously prevents any insight into the diastolic phase when LV pressure cannot be estimated. Nevertheless, it permits to obtain a measure of average end-systolic volume and pressure during the relatively long-lasting duration of acquisition. This low temporal resolution, coupled with the exclusion of extrasystolic and post-extrasystolic beats, offers the unique advantage of avoiding sudden variations that can hamper the analysis of a single cardiac cycle, as during echocardiography.
The decreased MMS was explained by an increased LV end-systolic volume much more than a decrease in systolic pressure. On one side, this finding confirms the notion that EVAR does not induce sudden shifts in arterial pressure. On the other hand, it indicates that the contractile impairment hampered the myocardial capability to decrease the LV volume against an unchanged aortic pressure. By contrast, the parallel response and variable of LV end-diastolic volume, at least partially prevented the capability of LVEF to describe this initial alteration. It is also necessary to consider that stiffness changes as represented in Figure 2 curves have not observed a unique trend between pre-operative and post-operative, although they have shown that a change is taking place probably influenced by the insertion of the endoprosthesis. The mechanisms underlying this contractile impairment were far beyond our clinical observational study. A possible direct interference of the afterload curve on cardiac performance might be further aggravated by the consequences of the altered sphygmic wave transmission on the driving pressure at the inlet of coronary arteries in the early phases of diastole. Reflected waves, coming first to the heart, also determine an increase of pressure in late systole, causing an increase in left ventricular filling and an increased oxygen consumption by the myocardium, which also contribute to further augment coronary risk.
Both factors might, indeed, contribute to trigger that series of events that has been found to explain the progressive decline in cardiac structure and function in patients with LV contractile impairment. 26

| Limitations
Our study has some limitations. The sample of patients analysed is small and without a control group. However, the data were prospectively collected, and we excluded those patients that were subjected to cardiac revascularisation to make the sample under analysis more homogeneous. This could also be linked to the analysis of the aortic stiffness, which results did not reach any statistical significance, even if a clear trend on parameters variation was observed. In fact, it is not in question the role that EVAR has on increasing aortic stiffness, but rather the clinical meaning in terms of cardiovascular complications. 10,21,27 These small variations between variables should also be observed considering the short temporal interval in analysis. Further studies with larger sample size and longer follow-up, including a control group, may help to validate the results and to evaluate the long-term impact of these variations.
In conclusion, the present data indicate that EVAR causes an altered transfer function of the sphygmic wave along the aorta. This pressure effect is early followed by a decrease in the MMS as an index of cardiac contractility. Despite the invariance of LVEF, this finding indicates the occurrence of an initial LV dysfunction. In these patients we could optimise drug therapy to improve heart remodelling during the follow-up. Although the evolution of this alteration cannot be defined based on the present data, these findings indicate that the hemodynamic effect of the applied aortic stiffness should be carefully evaluated to define candidates to EVAR as opposed to open surgery. Also, the implementation of endografts with less rigid and more adaptable geometry design could be helpful to overcome these issues.