One-year performance of biorestorative polymeric coronary bypass grafts in an ovine model: correlation between early biomechanics and late serial Quantitative Flow Ratio

Abstract

Open in new tabDownload slide

OBJECTIVES

This study aimed to investigate the impact of mechanical factors at baseline on the patency of a restorative conduit for coronary bypass grafts in an ovine model at serial follow-up up to 1 year.

METHODS

The analyses of 4 mechanical factors [i.e. bending angle, superficial wall strain and minimum and maximum endothelial shear stress (ESS)] were performed in 3D graft models reconstructed on baseline (1-month) angiograms frame by frame by a core laboratory blinded for the late follow-up. The late patency was documented by Quantitative Flow Ratio (QFR^®) that reflects the physiological status of the graft. The correlation between 4 mechanical factors and segmental QFR (△QFR) were analysed on 10 equal-length segments of each graft.

RESULTS

A total of 69 graft geometries of 7 animals were performed in the study. The highest △QFR at 12 months was colocalized in segments of the grafts with the largest bending angles at baseline. Higher △QFR at 3 months were both at the anastomotic ends and were colocalized with the highest superficial wall strain at baseline. High baseline ESS was topographically associated with higher △QFR at the latest follow-up. Correlations of minimum and maximum ESS with △QFR at 3 months were the strongest among these parameters (ρ = 0.30, 95% CI [−0.05 to 0.56] and ρ = 0.27, 95% CI [−0.05 to 0.54], respectively).

CONCLUSIONS

Despite the limited number of grafts, this study suggests an association between early abnormal mechanical factors and late flow metrics of the grafts. The understanding of the mechanical characteristics could help to improve this novel conduit.

INTRODUCTION

The inception of coronary artery bypass surgery was >5 decades ago, as one of the treatment potions for ischaemic heart diseases, especially for serious arterial disease. Intra-thoracic arteries, pedunculated or not, isolated radial arteries, gastroepiploic arteries, saphenous and tibial veins as well as synthetic grafts have been tested, adopted and sometimes abandoned [1]. Autologous vessels have limited availability and conduit length, may be of poor quality, harvesting them might be time-consuming and their harvesting may result in additional morbidity [2, 3]. Furthermore, the lack of available conduit was the reason for including 9.1% of the screened patients, who present with diabetes mellitus or possess a risk of wound complications, in the SYNTAX trial into a nested non-surgery (percutaneous coronary intervention) registry [4].

The perfect conduit for coronary artery bypass surgery would ideally exhibit high long-term patency, be readily available without complication arising from the traditional harvest of autologous graft and be available in sufficient length to revascularize all targets. Therefore, a restorative conduit, scalable in length and diameter to avoid a mismatch with the recipient native vessel and permanently available on the shelf, would be highly desirable as an alternative for autologous vessels. One of the major assets of the present polymeric conduit is its capacity to get biodegraded and to serve as a template for an indigenous, cellular and tissular restorative process that ultimately mimics the histological structure of native vessels. Previous preclinical and clinical experience with intracardiac bioresorbable polymeric conduits (Fontan tube and pulmonary conduits) and valvular leaflets (transcatheter aortic valve implantation) has demonstrated the feasibility of the concept [5].

Mechanical factors acting at the interface between graft and blood flow are believed to play a major role not only on the coverage of endothelial cells and neointimal hyperplasia but also on the late progressive atherosclerosis and thrombogenicity of neo-tissue [6]. For example, endothelial cells are very sensitive to mechanical factors (e.g. endothelial shear stress) and a wide variety of cell functions can be influenced through the activation of mechanosensitive receptors and signalling pathways [7]. Currently, several techniques of computational modelling based on angiograms are developed to quantify the mechanical factors or physiological function of vessels in vivo, such as dynamic superficial wall strain (SWS) [8], and endothelial shear stress (ESS) [9], as well as quantitative flow ratio (QFR) [10]. Therefore, investigation of the impact of these mechanical factors on the inter-lumen layer in vivo on the late patency might be helpful to optimal iteratively designing a novel biorestorative bypass graft.

In this study, we aim to investigate the potential association between mechanical factors in vivo at baseline (1 month) and haemodynamic performance at follow-up in ovine models implanted with this novel biorestorative graft. The baseline angiography was used to reconstruct 3D graft geometry and process these models with these computational techniques to assess bending angle, SWS and ESS and correlated these with segmental QFR (△QFR) at late follow-up.

MATERIALS AND METHODS

Ethics statement

Angiography data at baseline and follow-up were acquired from the animals implanted with the biorestorative bypass graft (Xeltis BV, Eindhoven, Netherlands). The study was conducted in accordance with the Guide for Care and Use of Laboratory Animals and the ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments) and was approved by the Test Facility’s Ethical Committee for compliance with regulations prior to study initiation (Protocols IQI001-IS02 and IQI005-IS02).

Study device

The biorestorative Xeltis coronary artery bypass graft (XABG) is composed of an electrospun supramolecular polymer that contained the ureidopyrimidinone supramolecular binding motif [11]. To prevent collapse and/or kinking of a bypass graft, micro-skeleton made of nitinol is encapsulated between the 2 layers of polymer and separated by a gap of 300 microns between each crown of nitinol. The length is 150 mm and the inner diameter is 4 mm.

Animal models

All animals were pre-treated with dual anti-platelet therapy of aspirin (325 mg on day 1, and 81 mg daily thereafter) and clopidogrel (150 mg on day 1, and 75 mg daily thereafter). On the day of operation, animals were anaesthetized, and a left lateral thoracotomy was performed to allow access to the heart and descending aorta. Left anterior descending (LAD) and descending aorta were isolated and assessed for appropriate distal and proximal anastomoses. Once heparin anticoagulation was initiated, cannulas were placed, and cardiopulmonary bypass was initiated. Cardiac arrest was induced with Plegisol or Del Nido, and additional cardioplegia solution was administered every 20 min as needed in order to prevent ischaemia.

The distal anastomosis was checked for hemostasis and clamped to prevent haemorrhage, while the aorta was punctured and anastomosed to the proximal end of the grafts. An arteriotomy was made in the partially occluded descending aorta using a stab incision and a 4–6-mm aortic punch to create the proximal anastomosis site. The graft was trimmed and sewn into the anastomosis. Prior to completion of the proximal anastomosis, air was evacuated from the graft by releasing the bulldog-type clamp, and the proximal anastomosis suture was tied tight after all air had escaped. The distal anastomosis was carefully crafted with a long orifice and a length that is between 2 and 3 times larger than the diameter of the LAD. Blood flow was established in the graft and any additional required repairs to the anastomoses were made. After graft implantation was completed, the LAD was ligated a few millimetres upstream of the graft distal anastomosis. The heart was defibrillated (if needed) to establish a sinus rhythm, and cardiopulmonary bypass was discontinued. Protamine sulphate was administered as needed to control any haemorrhage associated with the procedure. As a reference, saphenous vein grafts (SVGs) were implanted in 3 other sheep according to the local standards of care.

Workflow of the study

Figure 1 shows the workflow of the analysis regarding the relationship between the mechanical factors at baseline (1 month) and △QFR at follow-up (Supplementary Material, Appendix). Based on the facts that the restorative conduit exhibits a layer of neo-tissue already visible on the optical coherence tomography (OCT) images at 1 month follow-up, angiograms at this point of time were analysed and referred as baseline for investigation of the mechanical factors.

Angiography and three-dimensional reconstruction

Serial angiography was planned at 1, 3, 6, 9 and 12 months (Supplementary Material, Appendix). Three-dimensional reconstruction was performed using pairs of frames for the entire cardiac cycle. Post hoc synchronization between 2 projections was performed by ECG superimposed on the fluoroscopic images (n = 6) or by visual assessment at the time of early ventricular ejection (n = 1). All reconstructions geometry of grafts within cardiac cycle were used to analyse dynamic bending angle and SWS, while the geometry at end diastole was used to perform the ESS and QFR analysis.

Bending angle analysis

The bending angle was determined at sampling points along the centreline within 10 mm length, by calculating the tangential vectors from 1 point to the consecutive centreline points (Supplementary Material, Fig. S1).

Superficial wall strain analysis

The motion function of the grafts within a cardiac cycle was extracted from angiography, which is determined by the point-wise mapping relationship between 2 geometries at consecutive timepoints. The SWS was calculated by the variation of element lengths of meshes by an in-house algorithm (Supplementary Material, Appendix) [8, 12, 13].

Endothelial shear stress analysis

Direct in silico measurements of ESS on bypass grafts were carried out by using an in-house algorithm [9, 14] (Supplementary Material, Appendix). ESS was computed on all surface mesh elements as the product of fluid viscosity and near-wall velocity shear rate in the last cardiac cycle (at matching time point to angiographic frame) to derive a coloured ESS map. The minimum and maximum ESS were defined when the inlet velocity was minimum and maximum, respectively.

Quantitative flow ratio analysis

Vessel QFR was analysed from the proximal ostium of the graft, up to the first distal anatomical landmark in the native vessel (e.g. septal or first diagonal side branch). QFR analysis was conducted using the contrast flow model based on flow velocity computed from contrast bolus frame count [10] by QAngio XA 3D/QFR software (Medis Medical Imaging BV, Netherlands). △QFR was further derived by the difference between the values at the proximal and distal ends.

Statistical analysis

Initial descriptive analyses were performed, as appropriate, to summarize the distributions of physiological and mechanical parameters; the type of descriptive statistic(s) will be specified where reported in the text or tables. Where multiple measures per animal per visit are included for the outcome variable, and models were not explicitly used to account for within-animal, within-visit correlation, confidence intervals were generated using clustered bootstrap (resample by per animal). For descriptive purposes, Spearman’s rank correlation coefficient was estimated between each structural measure and the colocalized measurement in △QFR at follow-up.

Three linear, mixed-effects models including the mechanical factors at 1 month as fixed-effects were fitted involving: (i) no mechanical factors; (ii) mechanical factor averaged across the graft; and (iii) mechanical factors at corresponding colocalized segments. The adjusted R-squared from these models was used to compare the proportion of the variance in △QFR at late follow-up explained by theses combined 4 mechanical factors at baseline. P-values and 95% confidence intervals were not adjusted for multiple comparisons, and no a priori primary outcome was selected, nor was a statistical sample-size calculation performed. As such, type 1 and type 2 error were not strongly controlled for, and inferences drawn should be interpreted as exploratory rather than confirmatory (Supplementary Material, Appendix).

RESULTS

Fifteen animals were investigated for the primary end point in the preclinical study with XABG (n = 12) and SVG (n = 3) (Supplementary Material, Appendix). A total of 69 graft geometries within 1 complete cardiac cycle were reconstructed from baseline (1-month) angiography (Table 1). At 1-month follow-up, the minimum lumen diameter of the restorative conduits reduces to 2.66 ± 0.21 mm due to the formation of the layer of the neo-tissue, whereas the minimum lumen diameter of the SVG increases to 4.23 ± 0.24 mm due to slight expansion.

Characteristics of 4 mechanical factors at baseline

The segment-averaged maximum bending angles (mean ± standard deviation) were the highest at the distal anastomosis site (segment 10), followed by the middle part of the grafts (segment 4), and the lowest at the ostial site (segment 1): 19.36 ± 7.85°, 16.77 ± 5.44° and 8.65 ± 0.97°, respectively (Supplementary Material, Fig. S2). The segment-averaged SWS of grafts was significantly higher (P < 0.05) at the distal anastomotic sites (segment 10) and the aorto-ostial sites (segment 1) than in the middle part of the grafts (Fig. 2B). The segment-averaged minimum ESS at the distal anastomotic site (segment 10) was significantly higher than in the other segments; and the ESSs were similar in the other remaining segments (Fig. 2C). The segment-averaged maximum ESS at the distal anastomotic site (segment 10) was also significantly higher than in the other segments; whereas the ESS in the other segments fluctuated around 1.77–2.26 Pa (Fig. 2D).

A representative example (Fig. 3, right) shows that the highest SWS is located 105.1 mm from the ostium and is increased up to 0.71 in segment 8 at end-systole (timepoint 8). Noteworthy, in this particular case, the site of the highest SWS at baseline was co-localized with the beginning of the severe stenosis noted on the 6 months follow-up angiography (Fig. 4A2). The time-varying ESS shows that relatively high values of ESS occurred at mid-diastole near distal anastomosis sites due to mild relative narrowing (Fig. 3, middle, timepoints 2–5). Low ESS (<1 Pa) was found during other phases of the cardiac cycle. As expected, ESS was very low at the early-systole and a retrograde flow could be even documented (timepoint 9).

Mechanical factors at baseline and delta QFR at follow-up

The fitted average line of △QFR along the grafts at the latest follow-up has a similar pattern as the one of SWS at 1 month (Fig. 2B vs Fig. 2E). Table 2 summarizes the correlation between baseline mechanical factors and 3- and 6-month △QFR. Figure 5 shows the correlation between baseline mechanical factors and △QFR follow-up for the 10-segment level analysis. Based on the fitted linear mixed-effects models, the adjusted R-squared indicated a marginal increase in the explained variance of follow-up QFR by using localized, baseline mechanical factors in this manner, with a progression in R² from 0.12, to 0.14 to 0.16 with the inclusion of (i) no mechanical factors, (ii) mechanical factors averaged across the graft and (iii) mechanical factors at corresponding colocalized segments to follow-up △QFR measurements.

DISCUSSION

The long-term patency of synthetic biostable or bioresorbable coronary graft conduits implanted on a beating heart structure is a complex multifactorial process affected by various factors, involving biological and mechanical transductions. The present study focused on the mechanical factors and investigated the impact of local physical forces acting on a foreign conduit connecting the descending aorta and a ligated LAD artery of an ovine model. The findings of this study suggest that there is an association between mechanical factors at baseline and the localized alteration of QFR at follow-up.

The implanted grafts are subjected to 36 792 000 repeated systolic expansion and diastolic recoil per year (∼70 heart rate × 60 min × 24 h × 365 days). In addition to the cyclic variation of intravascular blood pressure, the synthetic conduit is stretched downwards by the systolic contraction of the left ventricle and the lowering of the diaphragm during the respiratory cycle, while the proximal anastomosis is subjected to the variation of the diameter and complex spatial motions of the descending aorta. Furthermore, multi-directional bending and stretching of the graft are unavoidable mechanical constraints.

The late patency of the graft at follow-up was assessed using QFR as this enables the colour-coded display of the pressure loss due to lumen narrowing observed during a pullback of a ‘virtual pressure wire’. That computed drop in pressure was segmented along the graft and expressed in △QFR in 10 segments of equal length. Thereby, co-localization and correlation between mechanical forces acting at 1 month and late localized change in QFR were depicted, which led to a better understanding of the association between the impact of mechanical factors and the late lumen alteration of the graft at that particular site.

The observations can be summarized as follows: (i) higher △QFR at 12 months were located in segments of the graft with the largest static and dynamic bending angles at baseline; (ii) higher △QFR were found at 3 months at both anastomotic ends of the grafts which were colocalized with the highest SWS documented at 1 month; and (iii) higher baseline ESS were in general associated with higher △QFR at latest follow-up.

The results of these dynamic morphological changes, captured in our frame-by-frame analysis, showed that the bending was minimal in the mid-distal part of the graft (segments 6 and 7, Fig. 2). In these segments, △QFR at the latest follow-up was not noticeable, suggesting that these segments are less subdue to the stress of bending. This may also be due to the uniform armoured structure in the straight graft segments, resulting in lower SWS and uniform distribution of ESS.

It has been reported that intimal hyperplasia (IH) is localized at the distal anastomosis and is less prominent at the ostium sites [15]. At 1 month, there was no pressure drop (△QFR) along all the segments of the graft except for the distal anastomosis exhibiting a very low △QFR (<0.01). This indicates that the grafts are widely patent and that their patency has not been altered by an angiographically detectable thrombotic and/or inflammatory process at that time. At the latest available follow-up, △QFR was the highest in the distal anastomotic segments, followed by the values observed at the ostium of the graft. Although both anastomosis could be smoothly transited between the biorestorative grafts and native vessels, relatively high SWS still occurs at both ends of the grafts (Fig. 2B). This could be explained that both ends of the grafts with unsupported nitinol rings are subjected to excessive local deformation due to the relatively low mechanical stiffness. It could trigger IH and result in stenosis in the long-term. Note that the presence of Nitinol micro-skeleton in the middle part of graft maintains the mechanical stability and relieve the high strain of graft. The Nitinol micro-skeleton, encapsulated between the 2 layers of polymer, will be eventually embedded within native tissue even when the polymer is fully absorbed and, thus, could reduce the risk of stent thrombosis by preventing any direct contact with blood flow.

In this study, time-varying ESS of the graft was analysed frame by frame, and the most prominent variations in ESS were observed at the sites of the distal anastomoses (Fig. 2C and D). It possibly was related to the diameter mismatch between the grafts and native coronary arteries. In particular, ESS at 1 month was correlated to the magnitude of segment-averaged △QFR. These results confirmed that the well-established fact that both high and low shear stress could cause IH [16, 17], which is suspected to be responsible for the alteration of the graft patency following its implantation [18].

Although the clustered bootstrap analysis showed that the confidence intervals of the estimated correlation coefficients are wide, there is some positive signal provided by these estimates. Furthermore, linear mixed-effects models, which included all available follow-up QFR and mechanical factors (Fig. 5), suggest that some additional information is provided by using localized mechanical factors in predicting localized △QFR compared to mechanical factors averaged across the graft.

Whether these findings could suggest potential ‘remedies’ in the design of the graft and/or in its surgical technique of implantation remains hypothetical. Thickening of the graft wall in the vicinity of the distal and proximal anastomotic sites might be considered, to render that part of the graft less sensitive to mechanical strain, bending and potential kinking. Minimizing the mismatch between the graft and the grafted native vessel is highly recommended considering the large variation of shear stress at the site of the distal anastomosis. In addition, a less angulated implantation of the graft on the native vessel with a shallow angle of implantation and with a large elliptical ostial attachment might be envisaged and has been advocated by some surgeons [19], which could mitigate or even abate flow disturbances, thereby diminishing undesirable flow separation at the anastomosis, and preventing a less adverse flow-related shear stress distribution. Obviously, in the human configuration the graft will have a vertical orientation from top to bottom, from the ascending aorta to the anterior wall of the heart: a topographical orientation diametrically at variance with our ovine model in which the graft has to cross the chest cavity in a horizontal trajectory between the posterior descending aorta and the anterior left ventricular wall. This will result in less bending compared to our ovine model in which the graft has to cross the chest cavity in a horizontal and curved trajectory between the posterior descending aorta and the anterior left ventricular wall. The mechanistic studies should be repeated on the angiographic analyses obtained from first-in-man implants of the novel XABG conduit.

Limitations

One of the limitations of the present study is the limited number of observations. This is due to the complexity of the surgical procedure and to the need for appropriately recorded angiographic views that allow for optimal three-dimensional reconstruction of the graft lumen. Cumulative tests and computational modelling based on animal models could be a useful tool for assessing its performance and identifying the causes of late patency. Although the results provided by the present study are hypothesis generating and exploratory, they yielded potentially useful clues on how to improve not only the experimental setting, but also the device and the technique of implantation in this proof of concept. It remains to determine whether the biological process will not be overwhelmingly dominant in the late fate of the graft. Postmortem histomorphometry will provide a single snapshot view of a complex and evolutive biological process that combines biodegradation with biorestoration.

CONCLUSION

This study demonstrated that early mechanical factors were associated with late localized alteration of QFR. Among these mechanical factors, the baseline ESS was the strongest parameter correlated with △QFR at 3 months. This dynamic interplay between mechanical factors and luminal surfaces sheds light on the biomechanical factors that might contribute to the late patency of this bioresorbed bypass.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Funding

This reseach was supported by the Zhejiang Provincial Natural Science Foundation of China under Grant No. LQ20H180004 and Science Foundation Ireland (15/RP/2765).

Conflict of interest: Eric K.W. Poon received supercomputer time from the National Computational Infrastructure LIEF Grant (LE190100021). Dr. Mohammed S. El-Kurdi, Martijn Cox and Jochen Reinöhl report to work as employee of Xeltis. Renu Virmani reports personal fees from Xeltis, Medtronic as well as Institutional grants from Abbott Vascular, Boston Scientific, Medtronic, Xeltis and Becton Dickinson. Dr. Yoshinobu Onuma reports institutional research grants related to his work as the chairman of cardiovascular imaging core labs of several clinical trials and registry sponsored by industry, for which they receive no direct compensation. Prof. Patrick W. Serruys reports personal fees from Sino Medical Sciences Technology, Philips/Volcano and Xeltis. The other authors have nothing to disclose.

Data Availability Statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions

Xinlei Wu: Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Validation; Visualization; Writing—original draft; Writing—review & editing. Masafumi Ono: Data curation; Formal analysis; Writing—review & editing. Eric K.W. Poon: Formal analysis; Writing—original draft. Neil O'Leary: Data curation; Writing—original draft. Ryo Torii: Investigation; Writing—review & editing. Johannes P. Janssen: Formal analysis; Writing—review & editing. Shuang Jie Zhu: Formal analysis; Writing—review & editing. Yves Vijgeboom: Formal analysis; Writing—review & editing. Mohammed S. El-Kurdi: Data curation; Investigation. Martijn Cox: Data curation; Writing—review & editing. Jochen Reinöhl: Data curation. Jouke Dijkstra: Writing—review & editing. Peter Barlis: Supervision. William Wijns: Supervision; Writing—review & editing. Johan H.C. Reiber: Supervision. Christos V. Bourantas: Supervision. Renu Virmani: Supervision; Writing—review & editing. Yoshinobu Onuma: Funding acquisition; Methodology; Supervision; Writing—review & editing. Patrick W. Serruys: Conceptualization; Funding acquisition; Methodology; Project administration; Supervision; Writing—original draft; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Vito Domenico Bruno, Dominique Shum-Tim, Antonio Miceli and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented at the 35th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 13–16 October 2021.

REFERENCES

ABBREVIATIONS

 
• ESS

  Endothelial shear stress

 
• IH

  Intimal hyperplasia

 
• LAD

  Left anterior descending

 
• QFR

  Quantitative flow ratio

 
• SVG

  Saphenous vein graft

 
• SWS

  Superficial wall strain

 
• XABG

  Xeltis coronary artery bypass graft

Author notes