EFFECT OF STENT-GRAFT COMPLIANCE ON HEMODYNAMICS AND AORTIC STIFFENING IN AN IN VIVO PORCINE STUDY

Abstract

Background:

Compliance mismatch is a key contributor to adverse vascular remodeling and long-term hemodynamic complications following endovascular aortic repair. However, few in vivo studies have systematically compared the biomechanical and hemodynamic impacts of compliant versus conventional stent-grafts.

Objectives:

In this study, we evaluated an elastomeric nanofibrillar stent-graft (NF-SG) against a commercially available stiff stent-graft (CS-SG) in a swine model, focusing on the effect of stent-graft type on the hemodynamic indices.

Methods:

Twenty Yucatan minipigs were divided into three groups: Control, CS-SG, and NF-SG. The latter two underwent endovascular implantation of the stent-grafts. Hemodynamic assessments were conducted at baseline, immediately post-implantation, and at 18-week follow-up. Local PWV, pressure, pulsatility, distensibility, and harmonic waveform analyses were performed at the ascending and descending thoracic aorta.

Results:

CS-SG implantation led to a significant increase in PWV (from 4.72 to 7.66 m/s), marked reductions in aortic pulsatility at the stented site (from 6.4% to 1.7%), as well as suppressed harmonic contribution (from 39.3% to 29.3%) and distortion (from 0.12 to 0.05), indicating impaired distal impedance regulation. In contrast, the NF-SG preserved baseline PWV (4.69 to 5.37 m/s) and maintained physiological waveform profiles, with minimal changes in harmonic contribution and distortion. NF-SG also showed a smaller reduction in pulsatility at the stented site (from 6.1% to 3.1%).

Conclusions:

These findings demonstrate that NF-SG exhibits superior biomechanical compatibility by preserving aortic compliance and normal hemodynamic function. Compliant stent-grafts may offer a promising strategy to reduce the long-term cardiovascular burden associated with conventional endovascular repairs.

1. INTRODUCTION

The aorta is the primary artery of the circulatory system, originating from the left ventricle of the heart and extending to the aortic bifurcation, where it divides into the right and left common iliac arteries. Its primary function is to transport oxygen-rich blood to organs and limbs. One of the aorta’s key biomechanical properties is its elasticity, which enables expansion and recoil during the cardiac cycle, ensuring efficient and continuous blood flow[1]. The aorta functions as an elastic reservoir, regulating blood flow through the Windkessel effect[2]. During systole, the left ventricle contracts and ejects blood into the aorta, causing its wall to expand and temporarily store blood and the energy of the cardiac ejection[3]. In diastole, as the left ventricle relaxes and refills, the aortic wall recoils, propelling the stored blood forward into the peripheral circulation[3] and perfusing the coronary arteries[1,4]. This mechanism plays a crucial role in maintaining continuous and stable blood flow to the organs throughout the cardiac cycle, while simultaneously reducing left ventricular afterload by buffering the systolic pressure (SP).

The aorta is susceptible to various diseases and injuries, which often necessitate urgent medical intervention. Some of the most common conditions include aortic aneurysm, aortic dissection, and blunt aortic injuries, which can occur in both thoracic and abdominal segments[5]. If left untreated, these conditions can severely impair aortic function, leading to life-threatening complications. In severe cases where medical therapies and other non-surgical approaches are insufficient, surgical intervention is required. Traditionally, open surgery was the standard approach for treating aortic diseases, involving a large incision in the chest or abdomen to replace the damaged aortic segment with a vascular graft[6,7]. However, in recent decades, minimally invasive endovascular aortic repair (EVAR) and/or thoracic endovascular aortic repair (TEVAR) have emerged as a standard treatment option for most aortic aneurysms, dissections, and trauma, offering lower short-term complication rates, shorter recovery times, and reduced mortality risk[8–10]. Endovascular technique involves inserting a stent-graft through a small incision in the groin or neck, and using guidewires and catheters to navigate and deploy the graft within the affected aortic segment. Once deployed, the stent-graft serves as a new conduit, reinforcing the weakened or injured aortic wall.

Despite superior short-term outcomes, long-term complications following stent-graft implantation have been reported. Stent-grafting has been associated with significant long-term alterations in the hemodynamic, biomechanical, and geometrical properties of the aorta and the left ventricle, particularly in younger patients[11–14]. One of the primary concerns is the stiffening of the native aorta with a stent-graft[3,11], which disrupts its Windkessel function by reducing the ability to expand and recoil during the cardiac cycle[15], ultimately leading to increased left ventricular load[16,17], hypertrophy[18,19], and other significant cardiac complications[20] reported in up to 40% of long-term EVAR survivors[21]. The diminished buffering capacity of the aorta leads to faster pulse wave propagation, causing higher systolic and pulse pressures, hypertension[11,12,17], and elevated biochemical markers, such as N-terminal pro B-type natriuretic peptide (NT-proBNP), indicating increased cardiac strain[22]. These complications significantly contribute to morbidity, mortality, and long-term healthcare costs in EVAR patients.

Notwithstanding the clinical significance of these long-term changes, current research on aortic and cardiac functional alterations following TEVAR/EVAR remains limited[8,23]. Although several in vivo[24,25], ex vivo[24,26], and in silico[24,27] studies have consistently reported increased segmental aortic stiffness due to stent-graft implantation, most investigations have focused primarily on immediate postoperative complications such as secondary interventions, endoleaks, and graft failures rather than more subtle long-term effects like aortic stiffening or left ventricular dysfunction[11]. With EVAR being widely used in older patients - who often have reduced cardiac reserve and are less able to tolerate the increased afterload associated with compliance mismatch[28] – and its growing application in younger trauma populations with more compliant aortas and longer life expectancy[17], understanding even subtle changes in aortic biomechanics and cardiac function, which may accelerate over time, has become increasingly important[29]. This shift underscores the need to develop physiologically compatible stent-grafts that minimize disruption to native vessel function. One of the primary challenges is the compliance mismatch between the stent-graft and the native aorta[4], which arises largely from the use of stiff graft fabrics such as expanded polytetrafluoroethylene (ePTFE) and polyethylene terephthalate (PET)[23]. Additionally, the metal skeleton design of conventional stent-grafts plays a contributory role in stiffening[30]. To mitigate this mismatch and preserve the Windkessel function, there is a need for developing next-generation stent-grafts that incorporate more compliant materials and optimized reinforcement structures capable of withstanding pulsatile loading while maintaining mechanical compatibility with the native aorta.

Several groups have been working on the development of next-generation aortic grafts and stent-grafts. Rovas et al. developed and optimized a multi-layer aortic graft designed to match native aortic compliance[31], and evaluated it in an acute porcine model, demonstrating superior distensibility and more physiological hemodynamics compared to standard PET grafts[32] immediately post-implantation. Kizilski et al. introduced a double-walled stent-graft design filled with gas between the layers, allowing for tunable compliance through volume modulation[33] and demonstrated reduced PWV compared to stiff stent-grafts in vitro. Singh and Wang introduced a bio-inspired stent-graft (CaT-SG) modeled after the caterpillar hydrostatic skeleton[34], which showed enhanced circumferential compliance, distensibility, and bending flexibility compared to commercial devices, however, without direct hemodynamic evaluation. Our team has also made notable progress in addressing these needs by developing an elastomeric nanofibrillar fabric engineered to replicate the mechanical behavior and microstructure of the healthy native aorta[35]. We have designed a stent-graft that integrates this nanofibrillar material with an optimized stent skeleton configuration and systematically evaluated the hemodynamic performance of this stent-graft in vitro[36]. The results demonstrated that hemodynamic disturbances, including changes in pulse wave velocity (PWV) and alterations in aortic compliance, were significantly lower with our stent-graft compared to conventional devices.

While these studies represent important progress toward addressing compliance mismatch, most remain limited in scope, experimental depth, or translational relevance. Many rely on computational or in vitro testing alone, without capturing the complex biological and physiological responses that occur in vivo. Others demonstrate promising mechanical innovations but lack direct hemodynamic evaluation, long-term follow-up, or systematic comparison between compliant and stiff stent-grafts using validated stiffness metrics. To address these limitations, the present work was designed to comprehensively assess the longer-term hemodynamic effects of our compliant stent-graft compared to a conventional stiff stent-graft using a physiologically relevant Yucatan miniature swine model, providing critical insights into potential clinical applicability.

2. METHODOLOGY

2.1. Study Design

Our study evaluated the immediate and longer-term hemodynamic effects of a commercially available stiff stent-graft (CS-SG) and a compliant stent-graft (NF-SG) in a large animal model. The primary objective was to investigate how stent-graft compliance influences metrics commonly used to evaluate aortic stiffness and overall hemodynamic adaptation over time. The timeline of the study, including surgical and imaging timepoints, is summarized in Figure 1.

Figure 1.

Experimental study design timeline illustrating key procedures and imaging time points across the 20-week study period.

All experimental animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Nebraska Medical Center and were conducted in compliance with PHS Policy, the USDA Animal Welfare Act, and AAALAC guidelines. The longitudinal nature of our study allowed for an evaluation of both immediate and longer-term effects of stent-graft-induced arterial stiffening and hemodynamic changes, enabling a detailed comparison of compliant vs. stiff endografts over time.

2.2. Animals

A total of 20 Yucatan miniature swine were randomly assigned to three experimental groups: a Control group (sham surgery, no stent-graft implantation, N=7), a CS-SG group (commercial stiff stent-graft, N=8), and an NF-SG group (our compliant stent-graft, N=5). At baseline, animals were 13.6 ± 2 months old and weighed 58.1 ± 5 kg. Yucatan miniature swine were chosen for their physiological similarity to humans, vessel sizes compatible with standard endovascular equipment, and limited growth potential. Older animals were used to further limit growth and because juvenile swine may respond differently to vascular injury[37].

At least one week before study initiation, animals were housed under controlled environmental conditions, ensuring consistent temperature, humidity, and light cycles, and were acclimated to minimize stress. The animals then underwent imaging and surgery, with all procedures performed under general anesthesia. Anesthesia was induced with an intramuscular injection of Telazol (4.4–5 mg/kg), Ketamine (2.2–2.5 mg/kg), and Xylazine (2.2–2.5 mg/kg), and maintained with 1–5% inhaled isoflurane via an endotracheal tube. To minimize thrombotic risk, all animals received oral aspirin (325 mg per day) and clopidogrel (75 mg per day) starting three days before the first surgery and continued daily postoperatively until sacrifice. Cefazolin (1 g) was administered prophylactically within an hour before the induction surgery to prevent infections.

2.3. Stent-Grafts

The stent-grafts used in this study are illustrated in Figure 2. The compliant stent-graft (NF-SG) was an in-house-developed device with a nanofibrillar fabric designed to closer match the natural compliance of the aorta compared with conventional stent-grafts. The NF-SG was constructed by integrating a tapered tubular elastomeric nanofibrillar graft (ENG) fabric with a nitinol skeleton, serving as a biomechanical alternative to traditional ePTFE and PET-based devices. The ENG graft was manufactured using electrospinning of biomedical-grade polyether urethanes (Pellethane^® 5863–82A and Pellethane^® 2363–55DE, Lubrizol, USA) dissolved in a dimethylformamide and tetrahydrofuran solvent mixture (Sigma-Aldrich, USA). The electrospinning process was conducted on a truncated conical mandrel using a DOXA Professional Lab Device 2.0 (Doxa Microfluidics, Spain), producing tubular grafts with an 18 mm to 14 mm diameters and 12.5 cm length (for further details on the fabrication process, refer to Maleckis et al.[35]). The ENG fabric had a circumferential stiffness of 1.6 MPa calculated as a slope of a stress-stretch curve in the physiological stress range of 75–100 kPa. In comparison, conventional graft materials had 9.5 MPa stiffness in this stress range, while mid-aged human aortas have 0.4–0.5 MPa stiffness. For detailed testing methodology and comparison plots, refer to Shahbad et al[36].

Figure 2.

Design features of our compliant stent-graft (NF-SG) and a conventional stiff stent-graft (CS-SG). (A) NF-SG configurations include a continuous spiral skeleton and a distinct ring skeleton, with detailed graft and Nitinol ring dimensions. (B) Commercial CS-SG (Gore Excluder AAA contralateral limb) used for comparison. L indicates graft or stent-graft length; Ø, diameter; t, strut thickness; l, strut amplitude length; n, number of struts per ring.

NF-SG incorporated two types of metal framework designs. Three out of the five NF-SGs had eight internal nitinol support rings (wire thickness: 0.30 mm and 0.36 mm; diameters: 22 mm, 24 mm, and 26 mm; strut length: 10 mm) spaced along the length, with 12 or 16 struts each. The remaining two NF-SGs featured a continuous spiral nitinol skeleton instead of rings, providing uniform radial support along the graft’s length while maintaining flexibility and compliance. Both designs included two bare nitinol anchoring rings (proximal: 0.46 mm wire thickness, 26 mm diameter, 17 mm strut length, 12 struts; distal: 0.46 mm wire thickness, 22 mm diameter, 10 mm strut length, 12 struts) positioned at each end, with one-third of the ring inside the graft and two-thirds outside for fixation. The final stent-graft dimensions, including bare stents, were 14.5 cm in length and 20 to 17 mm in diameter. The detailed design, dimensions, and flexibility of the NF-SG are shown in Figure 2A and B.

A conventional stent-graft (CS-SG) was a Gore Excluder AAA Contralateral limb endoprosthesis (W.L. Gore & Associates, USA) sized to the porcine aorta (Figure 2B). Two stent-grafts overlapping 2–3 cm were implanted in each CS-SG pig, covering the descending thoracic aorta distal to the left subclavian artery.

2.4. Imaging

Computerized tomography angiography (CTA) was performed at three key time points (Figure 1) to assess vessel dimensions, intervention success, and longer-term stent-graft performance. A baseline CTA at Week 0 determined aortic morphometry to guide stent-graft sizing. An intermediate CTA at Week 11 evaluated stent-graft positioning, early vascular adaptation, and potential complications such as migration. A final CTA at Week 19, just before the secondary (terminal) procedure, assessed longer-term arterial changes.

All animals were scanned in the supine position. Metoprolol (30 mg IV) was used when necessary to maintain a below 100 bpm heart rate (HR) during the scan. EKG-gated imaging was utilized to capture both systolic and diastolic phases. Imaging was done using a 256-slice GE Revolution CT scanner (GE Healthcare, Chicago, IL), producing 512 × 512 pixel images at a resolution of 0.23 mm with an axial slice thickness of 0.625 mm. Scans were performed at 120 kVp following IV injection of 100 mL of ISOVUE iopamidol contrast (Bracco Imaging, Milan, Italy). DICOM images were imported into Materialise Mimics v25 (Materialise, Leuven, Belgium) for 3D reconstruction using standard tools for masking, region growing, mask separation, and model generation.

2.5. Surgical Procedures

Two major surgical procedures were performed for all animals. At Week 2, an induction TEVAR surgery was conducted to implant the stent-grafts (for CS-SG and NF-SG groups) and to record intraoperative physiological measurements (aortic pressures, flow, and distensibility) before and after the stent-graft implantation. Control animals underwent a sham surgery that included catheterization and measurements identical to the treatment groups but without stent-graft implantation. At Week 20, a terminal surgery was performed to collect follow-up physiological data and euthanize the animals.

Induction surgery. Percutaneous access to the femoral vein and artery was established using micropuncture kits under ultrasound guidance. The 6 Fr arterial sheath was connected to a pressure transducer for continuous recording of the pressure waveform. Next, a pressure-flow ComboWire (Philips, USA) was advanced to the tip of the sheath, normalized to a pressure transducer, and placed in the descending thoracic aorta near the celiac artery for the duration of the surgery to provide continuous invasive blood pressure monitoring. The common carotid artery was percutaneously accessed, and an arterial sheath was placed. A second ComboWire was then advanced to record pressure and flow waveforms in several spots along the thoracic aorta, including the distal descending thoracic aorta near the celiac artery (DTA), the mid-ascending thoracic aorta (ATA), and the middle of the descending thoracic aorta, approximately at the level of the base of the heart (mDTA). Pressure transducer and ComboWires were connected to PowerLab 16/35 (ADInstruments, New Zealand) through ComboMap Pressure and Flow system (Philips, USA), and data were recorded with an acquisition rate of 1 kHz. Once pressure measurements were complete, an intravascular ultrasound (IVUS) catheter (Volcano Visions PV .035) was inserted through carotid artery access to record aortic pulsation at the same anatomical locations (ATA, mDTA, and DTA) over 10–15 cardiac cycles. Data were acquired with the Philips Core Mobile Imaging System with a sampling rate of 12 frames per second.

Stent-graft was sized using the baseline CTA imaging targeting 10% oversize. Stent-graft (either NF-SG or CS-SG) was deployed to the descending thoracic aorta with the proximal end just distal to the left subclavian artery. In humans, the aorta dilates to match the stent-graft diameter within 6–12 months[17], leading to a complete loss of pulsatility in the repaired segment. To accelerate this process and replicate the chronic luminal remodeling and stiffening in CS-SG animals, three additional balloon-expandable stents (Palmaz XL Transhepatic Biliary Stents (Cordis Corporation, USA) proximally and distally, and CP stent (B. Braun Interventional Systems Inc., USA) in the middle) were deployed. This approach ensured immediate expansion of the stent-graft to its final diameter, eliminating segmental pulsatility and reproducing, within the shorter study period, the loss of compliance typically seen in long-term human recipients.

Following stent-graft placement, both the pressure wire and the IVUS catheter were used again to measure arterial pressure and pulsatility at the same aortic locations. Access and measurement schematics are summarized in Figure 3. Once all measurements were completed, animals were gradually recovered from anesthesia and monitored until fully awake.

Figure 3.

Schematic of the surgical procedure and measurement setup. Stent-grafts were deployed in the descending thoracic aorta. Pressure and IVUS data were collected proximally to the stent-graft at the ascending thoracic aorta (ATA), mid stent-graft at the base of the heart (mDTA), and distally to the stent-graft at the descending thoracic aorta above the celiac artery (DTA) using carotid and femoral artery access. Representative pressure waveforms shown in this figure were measured during intraoperative recordings under anesthesia and ensemble-averaged over approximately 30 cardiac cycles.

Terminal surgery was performed similarly to the induction surgery to collect physiological data at the same aortic locations. Once completed, animals were euthanized by administration of cold cardioplegic solution to the heart.

2.6. Data Analysis

Intrinsic aortic stiffness cannot be directly measured in vivo; instead, it is commonly inferred from hemodynamic parameters derived from time- and frequency-domain analyses. Pulse wave velocity (PWV) and pulse pressure (PP) are widely used time-domain metrics, while harmonic contribution (HC) and harmonic distribution (HD) serve as frequency-domain indices of stiffness. In this study, we employed both time- and frequency-domain metrics to comprehensively assess stent-graft-induced aortic stiffening and resulting vascular changes following endovascular intervention.

2.6.1. Time Domain Metrics

PWV quantifies the speed at which pressure waves travel through the arterial system and is defined as $PWV=\frac{\mathrm{\Delta }x}{\mathrm{\Delta }t}$, where $\Delta x$ is the distance between two measurement locations along the artery, and $\Delta t$ is the time delay of the pressure wave between those points. While global PWV measurements, such as carotid-femoral PWV, are commonly used in in vivo studies, they provide an overall assessment of arterial stiffness rather than a localized evaluation[38]. In contrast, local PWV allows for segmental analysis, making it a more precise metric for assessing changes due to interventions such as stent-grafting[38,39]. In this study, local aortic PWV was measured as the time required for the pressure wave to propagate from mid ATA to the DTA. The distance along the centerline between these two points (from mid ATA to DTA) ($\Delta x$) was measured using 3D aortic reconstructions generated from CTA images before each surgery in Mimics software (Materialise, Belgium).

The pulse wave transit time ($\Delta t$) was derived from the pressure waveform analysis. To ensure accurate PWV calculation, a consistent reference point was selected from the pressure waveforms at the ATA and the DTA. This reference point was chosen from the systolic upstroke of the pressure waveform, which occurs before peak pressure, as it is less affected by reflected waves[38]. The 20% threshold algorithm was applied to determine this reference point. This method identifies the time at which the pressure reaches 20% of the total rise between diastolic pressure (DP) and SP, calculated as $P_{20\%}=\mathit{DP}+0.2\times \mathit{PP}$, where PP is defined as, $PP=SP-DP$. The time difference between the reference points at the ATA ($T_{20\mathrm{\%},ATA}$) and the DTA ($T_{20\mathrm{\%},DTA}$) was used to calculate the pulse wave transit time, defined as: $\mathrm{\Delta }t=T_{20\mathrm{\%},DTA}-T_{20\mathrm{\%},ATA}$_, (Figure 4C).

Figure 4.

Pressure signal processing and hemodynamic assessment. (A) Representative raw pressure waveform recorded from the ascending thoracic aorta (ATA), showing the selected time window centered around the systolic peak. (B) Ensemble-averaged pressure signal (bold black line) generated from multiple cardiac cycles (colored signals) to reduce noise and improve signal quality. (C) Left: Extraction of key pressure landmarks including systolic pressure (SP), diastolic pressure (DP), and the 20% upstroke reference point. Right: Transit time (Δt) calculation based on P _20% timing difference between ATA and the DTA pressure signals. (D) Frequency domain analysis of ATA and DTA pressure waveforms using Fast Fourier Transform (FFT), identifying the fundamental and non-fundamental harmonics and their amplitudes.

Pressure waveform processing, including ensemble averaging and extraction of key pressure landmarks, was performed automatically using a MATLAB script (The MathWorks, Inc., USA). Before extracting pressure landmarks, the continuous pressure waveforms recorded at the proximal (ATA) and distal segments of the stent-graft (DTA) were processed using ensemble averaging, generating a single representative pressure signal for each location. Ensemble averaging is a widely used and reliable technique for time-domain signal analysis, particularly when improving signal quality in physiological recordings. This method synchronizes and averages multiple repeated pressure waveforms, effectively reducing random noise and enhancing the signal-to-noise ratio (SNR)[40], thus improving the accuracy and reliability of subsequent calculations.

The SP was selected as the reference time point, ensuring alignment across all cardiac cycles. A time window was then defined around SP to extract a complete cardiac cycle for averaging (Figure 4A). This window was set as:

$$TimeWindow=[T_{\mathit{SP}}-X(ms),T_{\mathit{SP}}+X(ms\left)\right]$$ (Equation 1)

where $X\left(ms\right)$ represents a fraction of a second before and after SP, uniquely determined for each animal based on its HR during the recording. The time window was chosen to ensure that one full pressure cycle was captured, with additional milliseconds before and after the cycle to account for variations. To construct the ensemble-averaged signal, ~30 individual pressure waveforms were extracted within their respective time windows and then averaged at each corresponding time point across the entire window. This produced a single representative pressure waveform, minimizing transient fluctuations (Figure 4B). Once the ensemble-averaged pressure signal was generated, the MATLAB script automatically identified SP and DP and computed PP, mean arterial pressure $MAP=DP+1/3\times PP,P_{20\mathrm{\%}}$ and $\Delta t$ for each animal.

For additional analysis, PWV was also normalized to 80 mmHg to account for variations in pressure across animals. The detailed normalization method and corresponding results are presented in Appendix A.

In addition to pressure-derived metrics, aortic pulsatility and distensibility were quantified using IVUS-based diameter measurements to further assess local aortic stiffness. Pulsatility was calculated as the percent change in diameter between systole ($D_{\mathit{sys}}$) and diastole ($D_{\mathit{dia}}$), normalized to the diastolic diameter, using the following formula:

$$\mathit{Pulsatility}=\frac{D_{\mathit{sys}}-D_{\mathit{dia}}}{D_{\mathit{dia}}}\times 100\mathrm{\%}$$ (Equation 2)

This parameter reflects the vessel’s ability to expand with pulsatile pressure.

Diameter-based distensibility was calculated by further normalizing the pulsatility to the local PP, providing a size- and pressure-independent index of arterial stiffness suitable for comparison across animals and anatomical locations:

$$\mathit{Distensibility}=\frac{\left(D_{\mathit{sys}}-D_{\mathit{dia}}\right)/D_{\mathit{dia}}}{PP}$$ (Equation 3)

This metric removes the confounding effects of local pressure variability and is expressed in units of mmHg⁻¹.

To obtain $D_{\mathit{sys}}$ and $D_{\mathit{dia}}$, IVUS images were analyzed using MicroDicom software (Sofia, Bulgaria). For each animal and condition, measurements were performed over at least three full cardiac cycles to account for physiological fluctuations. Within each cycle, the maximum and minimum inner diameters were identified, and systolic and diastolic diameters were computed as the average of these values across cycles.

At the DTA and mDTA, where the cross-section was approximately circular (Figure 3), a circle was fitted in the lumen of each frame, and its diameter was recorded. At the ATA, where the catheter was often tilted with respect to the centerline and produced an oblique view of the cross-section (Figure 3), the minor axis of the resulting ellipse was used as the aortic diameter.

2.6.2. Frequency Domain Metrics

In addition to time-domain analysis, frequency-domain analysis of the arterial pressure waveform has been extensively used to assess vascular characteristics such as stiffness, impedance, and wave reflections[41]. Pathological conditions and vascular interventions can alter the shape of the pressure waveform, and frequency-domain analysis provides a powerful tool for quantifying these changes. Fast Fourier Transform (FFT) is a widely used method for converting pressure waveforms from the time domain to the frequency domain, allowing for the extraction of harmonic components that characterize the shape of the waveform[42].

The FFT decomposition of a pressure waveform results in a harmonic spectrum, which is unique for each waveform (Figure 4D). Each harmonic (H) corresponds to a specific wavelength and frequency, and its amplitude (A) represents the contribution of that frequency component to the overall signal[43]. The first harmonic (H₁), which corresponds to the HR frequency, always has the largest amplitude, making it the dominant contributor to the waveform. Subsequent harmonics (H₂, H₂, …, H₁₀) correspond to integer multiples of HR frequency (2HR, 3HR, …, 10HR), with their amplitudes representing the contribution of each harmonic to the overall waveform. In this study, harmonics beyond H₁₀ were considered negligible, as their contribution was insignificant and often indistinguishable from signal noise. The fundamental frequency was defined as H₁ (HR frequency), while H₂–H₁₀ were classified as non-fundamental frequencies.

A MATLAB script (The MathWorks, Inc., USA), utilizing MATLAB’s built-in FFT function, was developed to extract harmonic components from each pressure waveform. The amplitudes of harmonics H₁–H₁₀ were computed and denoted as A₁-A₁₀, providing a quantitative representation of pressure waveform composition (Figure 4D)[44].

Two harmonic-based indices, previously established in the literature, were used to analyze pressure waveform changes due to stent-grafting. Harmonic Contribution (HC) quantifies how total waveform energy is distributed between fundamental and non-fundamental frequencies, providing insight into whether vascular interventions alter harmonic balance. HC was computed as:

$$H{C}_{NF}=\frac{\sum _{n=2}^{10}{A}_n}{\sum _{n=1}^{10}{A}_n}\times 100$$ (Equation 4)

where $H{C}_{NF}$ represents the percentage contribution of non-fundamental harmonics (H₂–H₁₀) to total harmonic energy. The fundamental harmonic contribution $\left(H{C}_F\right)$ was then defined as:

$$H{C}_F=1-H{C}_{NF}$$ (Equation 5)

This metric aligns with previous studies that analyzed amplitude proportions for each harmonic separately to assess vascular stiffness alterations[43,44].

Harmonic Distortion (HD), introduced by Milkovich et al.[45], quantifies the degree of waveform distortion by measuring the power contribution of non-fundamental harmonics relative to the fundamental harmonic. It is defined as:

$$HD=\frac{\sum _{n=2}^{10}{\left|A_n\right|}^2}{{\left|A_1\right|}^2}$$ (Equation 6)

HD represents the ratio of non-fundamental harmonic power to fundamental harmonic power, serving as an indicator of waveform distortion. An ideal sinusoidal waveform has an HD value of 0, representing no distortion. Higher HD values (approaching 1) indicate an increased contribution of non-fundamental harmonics, reflecting greater waveform distortion. Changes in HD may result from either physiological elastic recoil or pathological alterations, depending on the context.

2.6.3. Statistical Analysis

Statistical analyses were conducted using RStudio to evaluate the effects of the intervention on changes in time-domain and frequency-domain stiffness metrics across three treatment groups (Control, CS-SG, and NF-SG) throughout the intervention period (Baseline, Immediate and Terminal). Shapiro–Wilk test was used to assess the normality of data distribution, while Levene’s test evaluated homogeneity of variances. Differences between groups were analyzed using one-way repeated measures ANOVA (RM ANOVA) and mixed-design ANOVAs. Mauchly’s test was applied to assess sphericity in RM ANOVAs, and when the sphericity assumption was violated, the Greenhouse-Geisser (GG) correction was applied. For post-hoc analysis, pairwise comparisons were conducted using Bonferroni correction to identify specific treatment groups and time points with statistically significant differences. A p-value < 0.05 was considered statistically significant for all analyses.

3. RESULTS

Tables 1 and 2 summarize the hemodynamic outcomes, presenting key stiffness-related metrics from both the time and frequency domains that were utilized to evaluate the effect of stent-grafting on pig hemodynamics and aortic biomechanics. These measurements were recorded at three time points: Baseline (pre-implantation), Immediate (immediately after implantation), and at the time of the terminal surgery (Terminal, 18 weeks after implantation).

Table 1.

Summary of heart rate and time domain stiffness-related metrics measured at the ascending thoracic aorta (ATA) and at the descending thoracic aorta above the celiac artery (DTA), and the middle of the descending thoracic aorta (mDTA) at three time points: Baseline, Immediate, and Terminal procedures.

Time Domain Stiffness Metrics	Treatment	Baseline	Time Point Immediate	Terminal
Heart Rate, HR (bmp)	Control	110 ± 17	NA	108 ± 29
	NF-SG	105 ± 22	93 ± 13	107 ± 26
	CS-SG	102 ± 15	113 ± 17	98 ± 10
PWV (m/s)	Control	4.8 ± 0.5	NA	5.2 ± 0.6
	NF-SG	4.7 ± 0.4	5.4 ± 0.5	5.4 ± 0.7
	CS-SG	4.7 ± 0.4	8.3 ± 1.9	7.7 ± 1.3
Pulse Pressure at ATA, PPATA (mmHg)	Control	24.5 ± 6.0	NA	23.6 ± 5.0
	NF-SG	22.5 ± 4.0	26.3 ± 4.5	28.5 ± 8.4
	CS-SG	24.1 ± 6.1	26.7 ± 5.2	26.7 ± 7.4
Pulse Pressure at DTA, PPDTA (mmHg)	Control	32.8 ± 6.0	NA	29.9 ± 6.8
	NF-SG	31.0 ± 6.0	22.0 ± 9.1	26.8 ± 3.7
	CS-SG	32.0 ± 6.3	26.3 ± 3.9	28.7 ± 8.0
Mean Arterial Pressure at ATA, MAPATA (mmHg)	Control	84.3 ± 19.4	NA	88.5 ± 23.7
	NF-SG	75.5 ± 18.2	69.4 ± 21.1	85.4 ± 23.6
	CS-SG	91.9 ± 21.6	76.4 ± 15.8	80.2 ± 23.2
Mean Arterial Pressure at DTA, MAPDTA (mmHg)	Control	86.7 ± 18.6	NA	96.3 ± 21.5
	NF-SG	73.9 ± 20.7	64.8 ± 29.9	86.7 ± 23.7
	CS-SG	97.4 ± 24.3	85.3 ± 18.3	90.3 ± 22.1
Pulsatility at ATA (%)	Control	11.7 ± 4.1	NA	12.4 ± 3.7
	NF-SG	11.4 ± 2.5	16.1 ± 4.6	14.0 ± 1.9
	CS-SG	11.3 ± 2.1	13.5 ± 3.5	12.6 ± 3.1
Distensibility at ATA (×10−3 1 /mmHg)	Control	4.9 ± 1.6	NA	5.3 ± 1.6
	NF-SG	5.1 ± 0.6	6.2 ± 1.8	5.2 ± 1.5
	CS-SG	4.9 ± 1.2	5.0 ± 0.8	5.0 ± 1.5
Pulsatility at DTA (%)	Control	4.7 ± 1.8	NA	5.5 ± 3.3
	NF-SG	3.7 ± 1.9	3.3 ± 1.1	3.9 ± 0.9
	CS-SG	5.1 ± 2.4	3.4 ± 2.1	3.4 ± 0.9
Distensibility at DTA (×10−3 1/mmHg)	Control	1.4 ± 0.4	NA	1.9 ± 1.0
	NF-SG	1.2 ± 0.6	1.7 ± 0.6	1.4 ± 0.3
	CS-SG	1.6 ± 0.7	1.3 ± 0.6	1.3 ± 0.4
Pulsatility at mDTA (%)	Control	6.1 ± 1.7	NA	5.5 ± 3.1
	NF-SG	6.1 ± 1.8	3.0 ± 1.4	3.1 ± 0.5
	CS-SG	6.4 ± 0.7	1.8 ± 0.6	1.7 ± 1.1
Distensibility at mDTA (×10−3 1/mmHg)	Control	2.3 ± 0.7	NA	2.1 ± 1.0
	NF-SG	2.4 ± 0.5	1.2 ± 0.6	1.5 ± 0.2
	CS-SG	2.5 ± 0.6	0.7 ± 0.2	0.7 ± 0.4

Values are presented as mean ± standard deviation for Control, NF-SG, and CS-SG groups.

Table 2.

Summary of frequency domain stiffness-related metrics measured at the ascending thoracic aorta (ATA) and the descending thoracic aorta above the celiac artery (DTA) at three time points: Baseline, Immediate, and Terminal procedures.

Frequency Domain Stiffness Metrics	Treatment	Baseline	Time Point Immediate	Terminal
Non-Fundamental Harmonic Contribution at ATA, HCNF, ATA (%)	Control	44.6 ± 7.7	NA	42.8 ± 10.1
	NF-SG	41.2 ± 6.9	42.4 ± 6.9	46.0 ± 10.3
	CS-SG	44.5 ± 9.3	46.5 ± 6.4	46.5 ± 6.5
Non-Fundamental Harmonic Contribution at DTA, HCNF, DTA (%)	Control	37.9 ± 4.0	NA	39.9 ± 6.0
	NF-SG	36.3 ± 5.1	36.8 ± 8.2	33.9 ± 4.5
	CS-SG	39.3 ± 5.2	28.5 ± 4.5	29.3 ± 3.3
Harmonic Distortion at ATA, HDATA	Control	0.18 ± 0.14	NA	0.17 ± 0.17
	NF-SG	0.12 ± 0.08	0.16 ± 0.07	0.26 ± 0.27
	CS-SG	0.18 ± 0.10	0.21 ± 0.10	0.22 ± 0.10
Harmonic Distortion at DTA, HDDTA	Control	0.10 ± 0.04	NA	0.11 ± 0.05
	NF-SG	0.08 ± 0.05	0.07 ± 0.04	0.07 ± 0.04
	CS-SG	0.12 ± 0.05	0.04 ± 0.03	0.05 ± 0.02

Values are presented as mean ± standard deviation for Control, NF-SG, and CS-SG groups.

3.1. PWV

PWV was derived from pulse transit time and centerline distance; therefore, regional differences in aortic geometry (i.e., cross-sectional area measurements) did not influence the measurements. Figure 5A illustrates changes in PWV for each treatment group (Control, NF-SG, and CS-SG) at baseline and terminal events, while Figure 5B depicts the longitudinal trend of PWV from baseline to terminal events in the NF-SG and CS-SG groups. At baseline, PWV was not different across all treatment groups (p = 0.821).

Figure 5.

Pulse Wave Velocity (PWV) changes across treatment groups and time points. (A) Box plots comparing PWV between Control, NF-SG, and CS-SG groups at Baseline and at the time of the Terminal evaluation. (B) Line plots depicting longitudinal changes in PWV for control, NF-SG and CS-SG groups across Baseline, Immediate, and Terminal events. While CS-SG implantation led to a significant and sustained increase in PWV, NF-SG did not induce significant changes, maintaining values comparable to baseline. ** denotes p<0.01, *** denotes p<0.001.

In the control group, PWV at the time of the terminal event was not significantly different from baseline (p = 0.389). In the CS-SG group, PWV increased by ~76% immediately post-implantation (p = 0.003) and was 68% higher at the time of the terminal procedure (p = 0.002) compared to the baseline. In contrast to CS-SG, NF-SG showed no significant effect of time point on PWV (p = 0.168).

PWV at the time of the terminal evaluation was significantly different across the three treatment groups (p < 0.001). Specifically, CS-SG resulted in a greater increase in PWV than both NF-SG (7.7 ± 1.3 vs. 5.4 ± 0.7 m/s, p = 0.002) and the control group (7.7 ± 1.3 vs. 5.2 ± 0.6 m/s, p < 0.001), while the difference between NF-SG and the control group was not statistically significant (p = 0.956).

3.2. Pulse and Mean Arterial Pressure (PP and MAP)

Figure 6 presents PP variations at two anatomical locations: ATA (proximal to the stent-graft) and DTA (distal to the stent-graft), across different treatment groups and time points. PP at the ATA location remained unchanged across all treatment groups (p = 0.646) and time points (p = 0.156). Both CS-SG and NF-SG groups showed no significant changes in PP at the ATA across time points, except for a significant increase in the NF-SG group between baseline and the immediate post-TEVAR measurement (p = 0.037).

Figure 6.

Pulse Pressure (PP) changes at proximal (ATA) and distal (DTA) locations across treatment groups and time points. (A, B) Box plots comparing PP at ATA (A) and DTA (B) between Control, NF-SG, and CS-SG groups at Baseline and at the time of the terminal event. (C, D) Line plots showing longitudinal changes in PP at ATA (C) and DTA (D) for control, NF-SG and CS-SG groups across Baseline, Immediate, and Terminal events. No significant differences were observed between treatment groups or time points, though both groups exhibited a slight upward trend at ATA PP, and transient reductions in PP at DTA after stent-grafting. * denotes p<0.05.

PP at DTA was not significantly affected by treatment type (p = 0.743) or time-points (p = 0.083. Both stent-grafted groups exhibited a non-significant pattern of lower PP immediately after stent-grafting, followed by a slight increase at the time of the terminal procedure.

MAP remained relatively unchanged across treatment groups and time points at both the ATA and DTA locations, with no statistically significant differences of the treatment type or time.

3.3. Aortic Pulsatility and Distensibility

The pulsatility and distensibility across different time points and treatment groups are shown in Figure 7C–D for the ATA location. All groups exhibited similar pulsatility at baseline (p = 0.977, Figure 7C). Mixed ANOVA analysis revealed no significant effect of treatment on ATA pulsatility (p = 0.591); however, a significant effect of time point was observed (p = 0.028), indicating that ATA pulsatility changed over time, likely in response to stent-grafting. NF-SG and CS-SG groups exhibited approximately 40% and 19% higher ATA pulsatility immediately after stent-grafting, respectively (p = 0.153 and p = 0.087). Although pulsatility declined by the time of the terminal event, the NF-SG group remained approximately 13% higher than the control group at that time point (p = 0.130). Distensibility, which normalizes pulsatility to local PP, is shown in Figure 7D. When corrected for PP, pulsatility changes after stent-graft implantation appeared reduced in the CS-SG group. ATA distensibility in the NF-SG group was slightly higher after stent-grafting, but this difference was not statistically significant (p = 0.388).

Figure 7.

Pulsatility and distensibility evaluated at proximal (ATA) and distal (DTA) aortic locations. (A, B) Representative IVUS cross-sectional images at systole and diastole showing diameter measurements at the ATA (A) and at the DTA (B). (C, D) Longitudinal changes in pulsatility (C) and distensibility (D) at the ATA across Control, NF-SG, and CS-SG groups. (E, F) Longitudinal changes in pulsatility (E) and distensibility (F) at the DTA for the same groups.

Figures 7E–F present pulsatility and distensibility measured at the DTA for all treatment groups and time points. Neither the between-group differences (p = 0.688) nor the time point effects (p = 0.133) were statistically significant in the DTA. Control and NF-SG groups showed similar mean pulsatility across time points. In the CS-SG group, pulsatility was 5.1% at baseline and 3.4% both immediately after stent-grafting (p = 0.124) and at the time of the terminal procedure (p = 0.091). To further evaluate these findings while accounting for pressure variability, distensibility was calculated and is presented in Figure 7F. No significant treatment (p = 0.859) or time point effects (p = 0.773) were found for the DTA.

Pulsatility and distensibility were also measured in the mDTA. This location was selected to assess how stented groups differed from the control group at the stented region of the aorta, with values reported in Table 1. Pulsatility changed significantly over time (p < 0.001) at this location, and these changes were influenced by treatment type (p = 0.033). At baseline, all groups showed similar pulsatility (p = 0.948); however, significant differences emerged after stent-grafting (p = 0.007), with the largest change observed in the CS-SG group. Implantation of the stiff stent-graft reduced vessel pulsatility from 6.4% to ~1.7% (p < 0.001). The NF-SG also reduced pulsatility, from 6.1% to ~3%, but these reductions were not statistically significant (p = 0.216 immediately post-implantation, p = 0.105 at the time of the terminal procedure).

After correcting for pulse pressure, distensibility showed a similar trend to pulsatility. Distensibility changed significantly over time (p < 0.001). At baseline, all groups had comparable values (p = 0.918), but at the time of the terminal surgery, significant differences were observed across groups (p = 0.004). In the CS-SG group, distensibility dropped markedly from 2.5×10⁻³ mmHg⁻¹ to 0.7×10⁻³ mmHg⁻¹ (p = 0.002 immediately post-implantation; p = 0.004 at the terminal event). The NF-SG group also experienced a reduction, from 2.4×10⁻³ to 1.2×10⁻³ and 1.5×10⁻³ mmHg⁻¹ at immediate and terminal time points, respectively, although the change was less pronounced and only significant at the time of the terminal procedure (p = 0.041).

3.4. Harmonic Indices

Figures 8A and 8C illustrate the HC and HD in the ATA across treatment groups (Control, NF-SG, and CS-SG) over time. Statistical analysis revealed no significant effect of treatment (p = 0.919) or time point (p = 0.314) on the proportion of fundamental vs. non-fundamental frequencies. Similarly, no significant changes were observed in HD between treatment groups (p = 0.814) or across time points (p = 0.552).

Figure 8.

Harmonic contribution (HC) and harmonic distortion (HD) at proximal (ATA) and distal (DTA) locations across treatment groups and time points. (A, B) Stacked bar plots comparing harmonic contribution at the ATA (A) and at the DTA (B) between Control, NF-SG, and CS-SG groups at baseline, immediate, and terminal time points. (C, D) Line plots showing longitudinal changes in harmonic distortion at the ATA (C) and DTA (D) for NF-SG and CS-SG groups. ** denotes p < 0.01

Figures 8B and 8D illustrate HC and HD in the DTA region across treatment groups and time points. Unlike the control and NF-SG groups, which remained relatively stable over time, the CS-SG showed a pronounced and progressive reduction in both non-fundamental frequency contribution (~25% drop) and HD (~50% drop). Statistical analysis revealed a significant effect of treatment (p = 0.029 for HC) and time point (p = 0.031 for HC, p = 0.050 for HD). At baseline, no significant differences were observed between groups (p = 0.558 for HC, p = 0.384 for HD), but at the time of the terminal surgery, CS-SG exhibited significantly lower values compared to the control group (p = 0.002 for contribution, p = 0.040 for distortion). Within CS-SG, both HC and HD dropped significantly from pre- to post-EVAR (p = 0.019, p = 0.007) and further at the time of the terminal event (p < 0.001, p = 0.002), reinforcing the substantial effect of CS-SG on harmonic wave composition. In contrast, the NF-SG and control groups exhibited minimal changes over time, indicating that the compliant stent-graft had a much weaker effect on harmonic parameters compared to CS-SG.

4. DISCUSSION

Compliance mismatch between the implanted stent-graft and the native aortic wall has long been recognized as a critical clinical concern, contributing to altered aortic hemodynamics and adverse vascular remodeling[46]. These differences primarily stem from the use of stiff graft materials, such as ePTFE and PET, as well as the rigid design of their metallic support structures[47,48]. While these effects may appear small in the short term, long-term complications are being increasingly reported and attributed to the biomechanical mismatch. Loss of aortic distensibility, reduced energy storage during systole, altered shape of the pressure waveform due to unnatural wave reflections, and increased PWV are among the key consequences[49]. Collectively, these changes contribute to elevated left ventricular load[16], increased myocardial mass[17], pathological hypertrophy[19], and, in severe cases, heightened risk of stroke and heart failure[21].

4.1. Contextualization of Key Findings

PWV, one of the most widely accepted surrogate markers of arterial stiffness, has been extensively investigated in the context of vascular interventions[50]. A broad consensus exists across in vivo, in vitro, and in silico studies that PWV significantly increases following EVAR and TEVAR, reflecting the stiffening effects of stent-grafts on the native aorta. For example, in older individuals, carotid-femoral PWV has been reported to increase from baseline values of approximately 12–13 m/s to 16–17 m/s six months after stent-grafting[22,51]. Similarly, brachial-ankle PWV measurements demonstrate the same trend, albeit with higher absolute values, rising from around 18 m/s at baseline to 19–20 m/s post-intervention, depending on the follow-up duration[18,52]. Notably, younger patients undergoing EVAR or TEVAR for blunt traumatic injuries have shown baseline PWV values in the range of 5–7 m/s, with post-intervention increases of up to ~40%, highlighting the disproportionate biomechanical impact of stiff stent-grafts in more compliant aortas[12,14,53]. This may appear to contradict recent reports that percent damping of diameter pulsatility is similar or lower in blunt thoracic aortic injury (BTAI) patients[54], but those measurements were taken ~1 month post-TEVAR - well before the 6–12 month adaptive dilation phase that abolishes segmental pulsatility. At this early stage, oversizing alone is insufficient to eliminate pulsatility in younger, more compliant aortas; indeed, the BTAI cohort was on average two decades younger than aneurysm and dissection patients and had higher baseline pulsatility. In contrast, older patients with stiffer aortas show smaller absolute post-TEVAR reductions but proportionally greater percent damping. Early hemodynamic factors, including hypotension and aortic contraction in acute trauma followed by pressure normalization at 1 month, may further affect these measurements.

Comparable trends have been observed in ex vivo models. In porcine aortic mock loop setups, segmental PWV was found to increase by ~5–25% after stent-graft deployment, depending on factors such as initial vessel stiffness, graft length, device type, and flow loop configuration[24,26,55]. Similar findings have been reported using silicone mock vessels, where baseline segmental PWV values increased from approximately 13 m/s to 16–18 m/s in stiffer models[56], and from 7 m/s to 9–10 m/s in more compliant models[36].

Consistent with previous studies on commercial stiff stent-grafts, the CS-SG group in our study exhibited a marked and sustained increase in PWV following implantation. Baseline PWV values were approximately 4.7 ± 0.4 m/s, comparable to those reported in porcine models and young human subjects[24,53,55], and increased by ~76% immediately after stent-grafting. Although a slight reduction was observed by the time of the terminal event, PWV (both before and after correction for BP) remained significantly elevated compared to baseline, indicating persistent arterial stiffening. In contrast, the NF-SG group showed only a modest, non-significant increase in PWV (~15%) immediately post-implantation, with values returning to baseline (~5.4 m/s, both BP-corrected and uncorrected) by the time of the terminal assessment. These results contrast with both our CS-SG cohort and the majority of previously reported stent-grafting outcomes, underscoring the NF-SG’s ability to preserve native aortic compliance and prevent intervention-induced stiffening.

PP was measured as a complementary metric to PWV to evaluate post-procedural hypertension and the physiological load on the aorta following endovascular repair [57]. Several in vivo studies have reported that PP increases following endovascular repair, with typical values rising from ~45 mmHg at baseline to ~60 mmHg post-intervention, and with 35% to 65% of patients developing elevated PP or systolic hypertension within five years post-implantation[12,14,17,58,59]. However, some studies have reported no significant change in PP even over longer follow-up periods[51–53]. These discrepancies may be attributed to variability in patient inclusion criteria, particularly the presence of pre-existing hypertension, as well as differences in implantation locations, measurement sites, and techniques. Ex vivo studies, which allow for localized pressure measurements at both proximal and distal sites of the stent-graft, have consistently shown an increase in PP at the proximal site (~15 mmHg), while distal PP responses have been more variable, ranging from mild increases to minimal or no change[16,36,55].

Our PP results at the ATA were consistent with previous studies. A slight increase in PP at the proximal site was observed in both stent-grafted groups compared to the untreated control group at the time of the terminal procedure (control: 23.6 mmHg; NF-SG: 28.5 mmHg; CS-SG: 26.7 mmHg). At the distal site of the stent-graft, the PP trend was reversed. Both stent-grafted groups showed a reduction in PP immediately following implantation, with a slight increase by the time of the terminal event. While none of these differences were statistically significant, the overall PP trend at the distal site appeared stable or mildly reduced, rather than elevated.

The typical physiological gradient in PP from proximal to distal segments was also affected by stent-grafting. At the baseline condition of all treatment groups, proximal PP was significantly lower than distal PP, reflecting the normal trend of increasing PP toward more distal muscular arteries[57]. After stent-graft implantation, however, this gradient diminished notably (CS-SG: +7.9 mmHg at baseline vs. +2.0 mmHg at terminal procedure; NF-SG: +8.5 mmHg at baseline vs. −1.7 mmHg at terminal event), indicating a flattening or reversal of the normal pressure profile along the aorta.

Aortic pulsatility and distensibility, as assessed by IVUS, serve as a localized marker of vascular stiffness and reflect the vessel’s ability to expand in response to pulsatile pressure. Unlike global metrics such as PWV, regional distensibility measurements can offer site-specific insight into mechanical behavior adjacent to the stented segment[60]. Although this area requires further study, previous investigations have generally reported reduced distensibility distal to the stented region[61], while proximal changes have been more variable, showing increases[62–64], decreases, or no significant change depending on the artery type and device design[48,65,66]. Our results align with these findings, particularly in the CS-SG group, where a marked drop in distensibility was observed distal to the graft, just above the celiac artery (DTA). In contrast, the NF-SG group showed only a minor decrease in distal distensibility immediately after stent-grafting, which was restored by the time of the terminal event. At the proximal site (ATA), changes in distensibility were relatively modest in both groups, although a slight increase was observed, suggesting preserved or improved local compliance.

While time-domain metrics like PWV, PP, and distensibility capture bulk changes in arterial stiffness, frequency-domain analysis in this study offered complementary insight into waveform morphology and wave reflection behavior, particularly through HD and HC measures. Prior studies have shown that an increased presence of non-fundamental higher-order frequency components in pressure signals can indicate enhanced wave reflections and altered impedance conditions[41,67]. Our study demonstrated notable shifts in both proximal and distal segments. Specifically, we observed that stent-grafting reversed the natural trends of HC and HD. In control animals, HC and HD values were lower proximally and higher distally at the time of the terminal event, reflecting the expected gradient of wave reflection in a healthy aorta. In contrast, both stent-grafted groups, particularly CS-SG, exhibited the opposite trend. This reversal was most pronounced at the distal site, where the CS-SG group demonstrated a significant drop in HC and HD. This reduction may result from proximal damping of high-frequency components by the stiff graft, as well as potential alterations in distal wave reflection caused by impedance mismatches.

4.2. Mechanistic and Clinical Interpretation

An increase in PWV has long been associated with arterial stiffening due to aging and stent-grafting, but the mechanisms behind this stiffening differ between age-related changes and stent-graft implantation. In aging, aortic stiffening results from degeneration of the medial layer. With advancing age, elastic structures gradually degrade and fragment, while stiffer collagen fibers accumulate and crosslink in the vessel wall[68]. This structural turnover reduces the aorta’s ability to stretch during systole and absorb the incoming pressure wave from the heart. As a result, the entire vessel becomes less distensible, limiting energy storage and accelerating forward wave transmission, ultimately leading to higher PWV[69]. In contrast, stent-graft-induced stiffening is localized to the region covered by the device and is externally imposed. A stiff stent-graft constrains the natural expansion and recoil of the aortic segment it covers, forming a rigid conduit with reduced capacity to absorb pulsatile energy[26]. Consequently, a larger portion of the stroke volume is transmitted downstream, increasing the velocity of wave propagation, consistent with the elevated PWV observed in the CS-SG group. The NF-SG, on the other hand, was specifically designed to preserve physiological distensibility. Its compliant materials and fiber architecture mimic the elastic properties of the native aorta, allowing the stent-grafted aortic segment to expand appreciably. This improved distension produces a larger diameter change during the cardiac cycle (dD), which, according to the Bramwell-Hill equation[70], reduces the pressure-to-diameter gradient (dD/dP) and slows the propagation of the forward wave. As a result, the NF-SG maintained PWV levels close to baseline and control values, supporting its role in mitigating the mechanical mismatch commonly introduced by conventional stent-grafts.

An increase in PWV alters the timing of pressure wave travel and reflection throughout the arterial system, which can significantly affect the shape and amplitude of the pressure waveform. When wave speed is elevated, the forward pressure wave reaches distal arteries earlier, and reflected waves return to the heart more quickly. As a result, these reflected waves arrive during systole rather than diastole and merge with the forward wave, leading to higher SP and lower DP. This widens the pressure range and increases PP[71]. However, the extent of this effect depends on the buffering capacity of the proximal aorta[72]. In natural aging where stiffening occurs throughout the aortic wall, including the ascending and arch segments, this buffering capacity is significantly diminished[73]. In contrast, in younger vessels, such as those used in our study, the proximal aorta remains structurally intact and rich in elastin. These segments retain the ability to expand during both systole and diastole, helping to absorb the augmented wave and mitigate the rise in PP. This could explain why the proximal PP increase observed in the CS-SG group was not statistically significant compared to the control group, and why no major difference was seen between CS-SG and NF-SG. These findings underscore the important role of proximal aortic compliance in buffering pressure wave interactions and modulating PP following stent-graft implantation. This interpretation is further supported by our IVUS-based distensibility measurements, which showed a slight increase in proximal distensibility (ATA) in both treatment groups, indicating preserved or even enhanced local compliance in the proximal segment. These observations are consistent with prior in vivo studies[52,59], which have reported inconsistent trends in post-procedural PP, likely reflecting differences in patient age, baseline vascular health, and proximal compliance. The more pronounced changes in proximal PP observed in vitro[36,55] may also be explained by this phenomenon. In those settings, silicone vessels lack the adaptive compliance of living tissue and, therefore, cannot respond flexibly to increased pressure loads, resulting in exaggerated pressure elevations following stent-graft deployment. NF-SG showed a greater proximal (ATA) pulsatility increase than CS-SG, contrary to expectations that stiffer grafts would augment upstream distension more. This may reflect NF-SG’s preserved compliance, which allows greater forward wave transmission and peak pulsatile expansion in the compliant ATA, whereas the rigidity of CS-SG physically limits local expansion despite generating stronger wave reflections. In healthy young pigs, the highly compliant ATA responds primarily to the forward-traveling wave from the left ventricle; with NF-SG, downstream compliance is preserved and reflections are smaller, so the forward wave arrives unimpeded and produces greater systolic expansion. In contrast, CS-SG markedly increases downstream stiffness, producing stronger reflections, but much of this reflected energy returns late in systole or during diastole, contributing little to peak ATA expansion.

Another contributing factor may be ventricular-vascular adaptation. Implantation of a stiff stent-grafted segment reduces overall aortic compliance, prompting the left ventricle to adapt by altering stroke volume and pressure generation to maintain forward flow. In the CS-SG group, elevated proximal PP may drive the uncorrected pulsatility increase, but once normalized to PP, the augmentation disappears - indicating a pressure-driven effect rather than a true increase in wall compliance. In contrast, with NF-SG, preserved segmental compliance may allow the LV and proximal aorta to maintain more physiological coupling, resulting in genuine increases in vessel wall motion rather than pressure-driven distension. In both groups, the increased normalized proximal distensibility observed immediately post-implantation returned to baseline at the time of the terminal event, suggesting adaptive remodeling of the proximal aorta over time.

Our findings from HC and HD analyses provide additional insight into the lack of significant PP differences in the proximal segments. In these regions, we observed a more pronounced role of non-fundamental frequency components and greater waveform distortion, particularly in the CS-SG group. These non-fundamental harmonics, which do not correspond to the HR frequency, are often associated with wave reflections originating from peripheral branches, impedance mismatches, or interactions with the stent-graft itself. The implantation of a stiff stent-graft may create a localized impedance mismatch, acting as a new site of wave reflection. When combined with existing reflections from arterial bifurcations and distal sites, this can result in highly variable and complex waveform patterns. Such variability in reflected wave timing and amplitude may blur the net hemodynamic effect on PP[57], making it more difficult to detect clear group-level differences. This is consistent with our observation that some animals, primarily in the CS-SG group, exhibited abnormally high or irregular PP values post-intervention, suggesting that individual differences in reflection dynamics may have played a significant role.

Although no significant differences in proximal PP were observed within the four-month follow-up, it is important to consider the potential long-term implications. In the presence of a stiff stent-graft, the ATA may need to perform additional work to absorb and buffer excess pressure, particularly when compensating for early-arriving reflected waves[72]. This sustained mechanical load may lead to structural fatigue, accelerating elastin fragmentation, medial degeneration, and premature stiffening of the proximal aortic wall - effects also observed in hypertensive patients[71]. Such progressive changes may explain why some in vivo studies report the development of systolic hypertension several years after stent-graft implantation[14,17,59]. While our study may not have captured these late-stage effects within the current follow-up window, it is plausible that differences in proximal PP, particularly between CS-SG and NF-SG, will become more distinct over longer periods.

In contrast to the proximal segments, we observed a decreasing trend in PP at distal sites following stent-graft implantation. Under normal physiological conditions, especially in young, healthy arteries, PP tends to increase along the length of the aorta[57]. This pattern results from natural impedance mismatches between the compliant proximal aorta and the stiffer distal muscular arteries, which promote wave reflection and pressure amplification in the periphery[57]. However, with aging and natural stiffening, the proximal aorta - being more compliant initially stiffens more rapidly than the distal arteries, leading to a reduction in the natural stiffness gradient along the aortic length[74,75]. As this gradient diminishes, a greater portion of the forward pressure wave is reflected earlier, before reaching the distal arteries, thus reducing the strength of reflection waves in downstream regions[76]. This results in more stable or even reduced PP in distal segments, as observed in both clinical and experimental studies[76,77]. Stent-grafting represents an exaggerated form of this same phenomenon. The implantation of a rigid graft in the thoracic aorta disrupts local compliance and creates a strong reflection site upstream of the distal vasculature, further flattening the stiffness gradient. As a result, a significant portion of the forward wave is reflected by the stent-graft itself, and fewer reflections are generated or transmitted from the periphery. These changes weaken wave reflections from distal sites, contributing to the lower PP observed in our study[76].

These interpretations were further supported by our harmonic analyses at the distal aorta site. In the CS-SG group, we observed a marked reduction in the contribution of non-fundamental frequency components, accompanied by lower waveform distortion. These changes indicate a disruption in normal wave reflection patterns and suggest that the stiffer stent-graft impaired distal impedance regulation more significantly than the compliant NF-SG. The reduced presence of non-fundamental harmonics in CS-SG animals likely reflects weakened reflections from the distal muscular arteries, consistent with the observed drop in PP. This was further confirmed by our IVUS findings, which showed a substantial drop in distal aortic distensibility in the CS-SG group. These mechanical changes likely stem from abrupt stiffness transitions at the graft-aorta interface, which reflect pressure waves and reduce wall motion downstream[48,61,65]. Such impaired reflection patterns can allow more forward pressure to reach distal vascular beds with less buffering, potentially increasing microvascular load and shear stress in high-flow organs such as the brain, myocardium, and kidneys[72]. In contrast, the NF-SG preserved a more physiologic harmonic profile at these sites, suggesting more balanced wave transmission and better protection of downstream hemodynamic stability.

These findings contrast with the small increases in distal PP commonly reported in vitro[55]. This discrepancy likely stems from the simplified architecture of mock flow loops, which typically represent the arterial tree as a single elastic tube with a single localized resistor to mimic peripheral impedance. While useful for basic hemodynamic characterization, these models fail to capture the complex, distributed nature of wave reflections in vivo. In the physiological arterial tree, impedance mismatches are not confined to a single location but are distributed across multiple branch points and vascular segments, leading to more nuanced wave propagation and reflection[78]. As a result, in vivo studies similar to this one can reveal pressure and harmonic changes in distal arteries that are not observed in simplified in vitro setups.

4.3. Study Limitations

Despite the strengths of this study, several limitations warrant acknowledgement. First, in this pilot proof-of-concept study, the follow-up period was limited to four months, which is likely insufficient to capture the full spectrum of vascular remodeling and long-term cardiovascular adaptations, particularly in the proximal aorta. Longer-term studies are needed to evaluate NF-SG performance and safety under clinically relevant conditions.

The second limitation is the use of healthy aortas rather than aneurysmal or otherwise pathological vessels. Although TEVAR is increasingly used for traumatic aortic injury in younger, otherwise healthy patients without pre-existing aortic disease, its primary indication remains the treatment of aneurysms. In such pathological settings, altered wall composition, geometry, and loading conditions may influence the biomechanical behavior and performance of compliant devices. Importantly, in patients with advanced vascular disease and compromised cardiac function, complete loss of the Windkessel effect may be especially harmful, potentially contributing to cardiac deaths reported after aneurysm repair with conventional stiff stent-grafts.

Third, our NF-SG group included only five animals, all with the same graft material but two different stent skeleton designs. Although we did not observe apparent differences in hemodynamic characteristics between the devices, the small sample size precludes device-specific sub-analyses. Next, although we measured markers of waveform composition such as harmonic contribution and distortion, a full wave separation analysis (e.g., using pressure-flow decomposition) would allow for a more direct assessment of forward and reflected wave interactions and energy transmission across the arterial tree. Moreover, the relatively low frame rate of IVUS acquisition (12 frames per second) may have limited the temporal resolution required to accurately capture the maximum systolic diameter, potentially affecting the precision of distensibility calculations. Furthermore, intra-subject variability in large-animal models such as pigs remains a challenge. Biological heterogeneity in growth rate, baseline vascular tone, and tissue remodeling capacity may introduce variability in hemodynamic responses. Although our findings were consistent across animals, a larger sample size would improve statistical power and further strengthen the reliability of observed trends. Additionally, anatomical differences between porcine and human aortas, particularly in arch geometry, may influence wave reflection patterns and limit direct extrapolation to clinical settings. Pigs have a “bovine-type” arch configuration without a separate brachiocephalic trunk, a variant also present in approximately 15–25% of humans[79,80]. In our study, however, stent-grafts were implanted distal to the left subclavian artery, making it unlikely that arch branching differences substantially affected the wave reflections at the stented segment. Finally, while the CS-SG design was reinforced to simulate full deployment and eliminate post-operative expansion effects, real-world variability in device positioning, oversizing, and deployment behavior remains an important consideration.

5. CONCLUSION

This study provides one of the first in vivo, longitudinal comparisons of a compliant stent-graft (NF-SG) and a conventional stiff stent-graft (CS-SG) with a focus on their impact on aortic hemodynamics. Using pressure-derived metrics such as PWV and PP, distensibility and harmonic analysis, we demonstrated that the NF-SG preserves more physiological mechanical behavior compared to the CS-SG. These results support the hypothesis that restoring or maintaining native aortic compliance can reduce the adverse hemodynamic consequences often associated with endovascular repair. Importantly, the integration of advanced harmonic analysis allowed us to move beyond traditional stiffness metrics, capturing the subtler effects of wave reflections and frequency distortion along the aorta. The combination of the Yucatan minipig model and precise, site-specific pressure recordings provides a robust platform for evaluating the biomechanical implications of endovascular devices.

Abbreviations and Acronyms

EVAR

Endovascular Aortic Repair

TEVAR

Thoracic Endovascular Aortic Repair

BTAI

Blunt Thoracic Aortic Injury

SP

Systolic Pressure

DP

Diastolic Pressure

PP

Pulse Pressure

BP

Blood Pressure

MAP

Mean Arterial Pressure

e/PTFE

expanded/Polytetrafluoroethylene

PET

Polyethylene Terephthalate

PWV

Pulse Wave Velocity

CS-SG

Conventional Stiff Stent-Graft

NF-SG

Nano-Fiber Stent-Graft

CTA

Computerized Tomography Angiography

HR

Heart Rate

ATA

Ascending Thoracic Aorta

DTA

Distal Descending Thoracic Aorta

mDTA

Middle of Descending Thoracic Aorta

IVUS

Intravascular Ultrasound

HC

Harmonic Contribution

HD

Harmonic Distortion

FFT

Fast Fourier Transform

APPENDIX A. NORMALIZED PWV ANALYSIS

PWV is influenced by multiple factors, with age and blood pressure (BP) being the most significant[81]. Numerous studies, particularly those investigating hypertension-induced arterial stiffness, have demonstrated a strong correlation between PWV and BP parameters[19], including SP, DP, and PP. This dependency can lead to biased estimates when evaluating the effects of various treatments or interventions on PWV, particularly in studies where BP is also affected, such as this one. To isolate the intrinsic effect of stent-grafting on arterial stiffness, it is necessary to account for variations in BP when interpreting PWV measurements. Several statistical, mathematical, and empirical correction techniques have been proposed to normalize PWV for BP, with the choice of the method depending on the study design[82]. In this study, we employed a mathematical model proposed by Giudici et al.[83], which is derived from the Bramwell-Hill equation and incorporates the nonlinear pressure-area relationship for arteries.

According to this model, PWV at a desired reference pressure ($P_T$) can be calculated using the experimentally measured PWV $\mathrm{P}\mathrm{W}\mathrm{V}\left(PWV\left(P_C\right)\right)$ at the pressure $\left(P_C\right)$ recorded during the experiment, as follows:

$$PWV\left(P_T\right)=\sqrt{PWV{\left(P_C\right)}^2\frac{P_T}{P_C}+\frac{P_T}{\rho }\mathrm{l}\mathrm{n}(\frac{P_T}{P_C})}$$ (Equation A.1)

where, $PWV\left(P_T\right)$ represents the BP-normalized PWV at the target pressure $P_T$, $PWV\left(P_C\right)$ is the experimentally measured PWV at a given $P_C$, $\rho =1060\mathit{k}\mathit{g}/{\mathit{m}}^3$ is the density of blood, and $P_C$ is the BP at which the experimental PWV was recorded.

Since $P_C$ varies based on the PWV estimation method[83], we used the pressure at 20% of the systolic upstroke ($P_{20\mathit{\%}}$) as $P_C$. The target pressure ($P_T$) was set to 80 mmHg, which was determined as the average $P_{20\mathit{\%}}$ across all pigs and procedures. Equation A.1 allowed us to convert PWV values measured at varying experimental pressures ($P_C$) to a standardized pressure of 80 mmHg $\left(PW{V}_{@80}\right)$, enabling a more accurate comparison of aortic stiffness across different conditions. Table A.1 summarizes BP-corrected PWV outcomes.

Table A. 1.

BP-Corrected PWV (PWV@80) for the treatment groups at three time points: Baseline, Immediate, and Terminal.

Time Domain Stiffness Metrics	Treatment	Baseline	Time Point Immediate	Terminal
BP-Corrected PWV (m/s)	Control	4.8 ± 0.7	NA	5.1 ± 1.0
	NF-SG	5.1 ± 0.6	6.3 ± 1.0	5.4 ± 1.2
	CS-SG	4.4 ± 1.1	8.9 ± 2.3	8.1 ± 2.1

Values are presented as mean ± standard deviation for Control, NF-SG, and CS-SG groups.

When PWV was corrected for $P_T$ (i.e. 80 mmHg), a similar overall trend was observed compared to unadjusted PWV (Figure A.1A, A.1B). At Baseline, $PW{V}_{@80}$ did not differ significantly across groups (p = 0.405), confirming that animals had comparable vascular and hemodynamic conditions prior to intervention. Following stent-grafting, distinct responses were observed across treatment groups. In the control group, $PW{V}_{@80}$ remained unchanged throughout the study period (p = 0.909), indicating that no significant intervention-related effects were present. In contrast, treated pigs exhibited an increase in $PW{V}_{@80}$, immediately after implantation and over time. In the CS-SG group, $PW{V}_{@80}$ increased significantly from 4.4 ± 1.1 m/s at baseline to 8.9 ± 2.3 m/s immediately post-stent-grafting (p < 0.001). Although a slight reduction was observed at the time of the terminal event (8.1 ± 2.1 m/s), $PW{V}_{@80}$ remained significantly elevated compared to baseline (p = 0.012). This suggests that despite some level of vascular adaptation, arterial stiffening persisted following CS-SG implantation. In contrast, the NF-SG group exhibited a more stable $PW{V}_{@80}$ profile. $PW{V}_{@80}$ was approximately 23% higher immediately post-stent-grafting, but this difference was not statistically significant (p = 0.132). Furthermore, by the terminal procedure, $PW{V}_{@80}$ returned to baseline levels (5.4 ± 1.2 m/s, p = 0.319), reinforcing the finding that NF-SG does not induce sustained arterial stiffening over time. Comparison of $PW{V}_{@80}$ across treatment groups at the terminal procedure revealed a significant overall difference (p = 0.006), highlighting that stent-graft characteristics influence aortic mechanics. Specifically, CS-SG resulted in significantly higher $PW{V}_{@80}$ than the unstented control group (p = 0.009). $PW{V}_{@80}$ in CS-SG was 2.7 m/s higher than NF-SG (p = 0.027).

Figure A. 1.

BP-Corrected PWV (PWV@80) changes across treatment groups and time points. (A) Box plots comparing PWV@80 between Control, NF-SG, and CS-SG groups at Baseline and Terminal time points. (B) Line plots depicting longitudinal changes in PWV@80 for control, NF-SG and CS-SG groups across Baseline, Immediate, and Terminal timepoints. While CS-SG implantation led to a significant and sustained increase in PWV@80, NF-SG did not induce significant changes, maintaining values comparable to baseline. * denotes p<0.05, ** denotes p<0.01, *** denotes p<0.001.