Title: Comparative Review of Large Animal Models for Suitability of Proximal Aortic Endovascular Repair

word count: 250 3 Number of Figures and Tables: 2 Figures and 3 Tables 4 5 Personal, Professional, and Institutional Social Network accounts. 6

animal models to better contemporary understanding of endovascularly treating ascending ADs, also known as 6 Stanford Type-A ADs. This narrative review provides a current literature summary on this topic, including the 7 gross anatomical differences between adult porcine, ovine and bovine species versus that of their human 8 counterparts, as well as specific valvular and coronary vasculature measurement variances. To achieve this, 9 an electronic search of Cochrane Library, PubMed and Ovid Medline databases from January 1965 to June 10 2020 was performed, limited to articles published in English. A total of twenty-three research papers were 11 included in constructing this review. Our findings showed that whilst macroscopic anatomy remains grossly 12 similar, differences in valvular leaflet shape are present, with porcine and ovine models possessing anatomic 13 characteristics that are comparable to their human counterparts. Research into inter-species ascending aortic 14 anatomy has not been extensively performed, highlighting a literature gap. Conversely, multiple morphological 15 studies have highlighted that porcine coronary vasculature is closely resemblant to that of humans. In 16 conclusion, both porcine and ovine species are suitable as appropriate animal models in examining the 17 feasibility of endovascular stent-grafts for ascending ADs. However, given the similarities in coronary and aortic 18 valve anatomy to their human analogues, porcine models are better suited for this purpose. IJMS

INTRODUCTION.
1 The usage of non-human tissue in cardiothoracic medical research has markedly risen over the last half a 2 century, as a solution to both the ethical dilemmas posed by using, as well as the lack of readily available 3 human tissue for creating experimental clinical models. 1 One example of research involving such animal 4 models is seen in better understanding treatment outcomes for acute aortic dissections (AD), a life-5 threatening pathology that carries significant mortality rates of over 70% within one week of onset when left 6 untreated. 2,3 Several classifications of ADs currently exist, but arguably perhaps, one of the most commonly 7 used is the Stanford classification system. This system categorizes dissections on the basis of site of 8 intimomedial tear as either Type-A, defined as any AD involving the ascending aorta or Type-B, which are 9 ADs not involving the ascending aorta (NB. This review focuses primarily on Type-A ADs). 4

11
With few exceptions, the management of acute Type-A ADs is touted as a surgical emergency. 5,6 Given the 12 aforementioned high rates of mortality otherwise, there are few reasons for not following through with 13 operative treatment of Type-A ADs, with the main cited reasons being presence of significant medical 14 comorbidities that impact survival to one year or less, as with very advanced age and frailty, advanced 15 malignancies or patient/family wishes. 7 The surgical intervention for Type-A ADs has seen a marked evolution 16 over the years, due to the intertwined combination of technological improvements in equipment, as well as a better understanding of its natural history. At present, open surgical repair (OSR) remains the gold standard of (TEVAR) has heralded a paradigm shift in treatment options for aortic disease involving the descending aorta, 20 and as such has been viewed as a potential option for ascending aortic repair, and consequently Type-A AD 21 surgical repair as well. 9 As a result, selected patients who would otherwise be ineligible for OSR as aforementioned, which typically comprise up to 20% of all individuals, would benefit from having the 23 opportunity of still receiving life-saving treatment in the form of minimally invasive endovascular techniques. 10

25
There are various types of endovascular therapies currently viewed as a potential solution in treating Type-A 26 ADs, including branched stent-grafts and valve-carrying conduits. 10 However, the employment of these novel 27 therapeutic procedures within a clinical setting remains limited, with isolated case-reports and case-series 28 providing the bulk of currently available literature on patient outcomes. Consequently, there exists an urgent 29 requirement to choose appropriate animal models in order to better our understanding of endovascularly 30 treating Type-A ADs.

32
While there is a widespread amount of published research on the variances of cardiothoracic anatomy in non-33 human species, there exists no literature review synthesizing this information, highlighting the accelerated 34 need for one to be formulated. Consequently, this review article aims to combat this issue by providing a 35 summary of currently available information on this topic, with a particular focus on determining which animal 36 model amongst those of adult porcine, ovine or bovine species would be ideal for research pertaining to 37 endovascularly treating Type-A ADs, relevant to the practising surgeon. Three broad sections shall be ovines and bovines. The review shall then focus on specific aspects of cardiothoracic anatomy, explicating in IJMS would be best for being used as clinical experimental models, from a strictly anatomical standpoint for 1 bettering our understanding of Type-A AD treatment shall then be explored. IJMS

1
To appropriately answer the aforementioned questions on this topic, two main databases were utilized. These

6
Within Ovid Medline, since the term 'Type A aortic dissection' is quite well known within medical literature (as 7 opposed to its verbatim analogue 'Stanford Type A aortic dissection'), the search string was commenced by 8 initially mapping the keyword 'Endovascular' with the MeSH term 'Type A aortic dissection'. This was followed 9 by using the Boolean operator 'AND'. The keyword 'models' was them utilized, and finally, the Boolean

17
Finally, to obtain a better pictorial representation of the cardiothoracic anatomical variations between porcine, 18 ovine and bovine models, images from the University of Minnesota Atlas of Human Cardiac Anatomy were 8 IJMS humans 3 In spite of the advantages the utilisation of thoracic endovascular aortic repair (TEVAR) affords, including 4 eliminating the needed for perioperative cardiopulmonary bypass and the requirement for a major operative 5 incision such as a sternotomy, there exist certain limitations that prevent its routine employment in currently 6 treating Type-A ADs. 4,11-13 Given the paucity of large-scale trials documenting its efficacy however, as well as 7 long-term follow-up of patients who receive this modality of treatment, there exists a literature gap in 8 describing the specific limitations of endovascular therapy for ascending aortic pathologies. That being said, 9 the anatomical constraints of this novel therapy have received particular scrutinization, and shall now be 10 explored further.

12
One of the major challenges in successfully treating Type-A ADs with currently available stent-grafts lies in the 13 need to insert a straight device into a curved structure i.e. the aortic arch, which poses a high risk of 14 developing an endoleak. In an attempt to simplify landmarks within the complex anatomy of the aortic arch, 15 the Ishimaru classification is commonly used to categorize thoracic aortic 'zones' for stent-grafts. 14

17
Utilising Ishimaru's zone classifications, it is essential to ensure a 'safe' distance between the proximal and aortic rupture. 3,15,16 However, this measurement remains dependent on both the characteristics of the chosen 20 stent-graft, as well as the technical expertise of the operating doctor. Consequently, although there exists 21 some variation in what constitutes a 'safe' distance, a proposed criterion has been a length of at least 20 22 millimetres between the two landing zones, to avoid aortic rupture during graft deployment. 16

24
Secondly, problems are also created by the entry dissection tear occurring proximally within Zone 0 as 25 illustrated in Figure 1, specifically proximal to the sinotubular junction. A tear occurring within this region 26 would fail to allow endograft deployment in a manner that would allow coronary blood flow to be maintained. 15

27
Occlusion of the coronary ostia by closed ends of the stent-graft would cause ischaemia of the myocardium, 28 resulting in potential irreversible damage. 17, 18 Additionally, those with Type-A ADs extending into the aortic 29 valve would not be suitable for endovascular treatment with conventional stent-grafts, a situation typically 30 seen in between 10-20% of patients. 15 This is because at deployment the tip of the device must cross the 31 aortic valve, which could eventuate in possible ventricular perforation. Although this would pose a barrier to 32 treatment with currently available stent-grafts, given that they possess a distal cone that prevents their 33 deployment too close to the aortic valve, a proposed method to combat this has been suggested in the form of 34 novel 'valve-carrying conduits'.

36
Thirdly, variations with the anatomy of the normal aorta may interfere with a wholly endovascular modality of treatment for Type-A ADs. For instance, in those who have received prior coronary artery bypass surgery, the 38 presence of coronary grafts arising directly from the ascending aorta would present an increased risk of 39 myocardial ischaemia during endograft deployment. 15,16 IJMS 1 Based on these caveats, it is evident that the anatomy of the ascending aorta, aortic valve and coronary 2 vasculature are of particular significance in determining an appropriate animal model for Type-A dissection 3 research, which shall be addressed in the following sub-section. IJMS

Introduction and general cardiac anatomy 1
Similar to those of humans, the holistic cardiac anatomy of large mammals is analogous; four cardiac valves 2 are present with similar structures comparable to most quadruped mammals. Whilst human hearts can appear 3 in a variety of shapes, including elliptical, trapezoidal, and 'valentine', porcine species tend to be valentine 4 shaped, whilst the ovine heart varies from valentine to conical in shape, as illustrated in Figure 2. 19

6
With respect to the hearts of porcine and ovine species, the distance between the posteroinferior base to 7 apex, left lateral base to apex and the length of coronary sinuses are all significantly greater in their human 8 counterparts. As expected therefore, in conjunction with its comparatively larger size, the average human 9 heart maintains a larger organ to body weight ratio to that of both porcine and ovine species. A similar 10 scenario is visible in that of bovines, which possesses a nearly identical organ to body weight ratio of the

30
Differences in aortic valve leaflet shape and structure are also present, with only porcine valve leaflet depths 31 being comparable to their human analogues, although specimen analysis visualized more inter-species 32 variation between individual leaflets in the former. 20 Variations in leaflet thickness are particularly important to 33 make note of, as thin and fragile leaflets such as those seen in ovine species may not be structurally strong 34 enough to support heavy pressure loads during clinical usage for long periods of time.

Aortic anatomy 37
Unlike the aforementioned aspects of valvular anatomy, research into specifically the ascending aortic 38 differences between human and non-human species has not been extensively performed, highlighting a IJMS characteristics of the largest artery in mammals have been documented. Primarily, compared to the human 1 heart, the porcine species has only 2 head branches that originate from the aortic arch.

3
Dimensionally, the diameter of the proximal aorta amongst porcine species at its largest part is about 21% 4 lesser than that of their human analogues. What is noteworthy is that unlike their human counterparts, which 5 exhibit a gradual diameter decrease in tapering fashion, there is a sharp decrease in aortic diameter from the 6 descending thoracic aorta to the abdominal aorta within porcine, with exact values having been documented 7 in Table 2. Conversely, whilst research on the aortic anatomy of ovine species is scarce, it is known that the 8 ascending aorta itself, whilst maintaining a similar aortic diameter to that of their human counterparts after 9 accounting for the changes in organ to body weight ratio, is relatively short, of which the implications shall be 10 discussed in the next section. 27 There is also a marked decrease in the number of elastic lamellae within 11 ovine aorta, greatly reducing its mobility as well. 27

13
Finally, of the three non-human species described in this paper, perhaps the most review has been done on 14 bovine ascending aortic anatomy, with the 'bovine aortic arch' having been described as the single most 15 common congenital aortic anatomic variant within humans as well. Whilst this term itself is a misnomer, it is 16 used to supposedly refer to the variant within bovine species, which is characterised by a common single brachiocephalic trunk trifurcating into bilateral subclavian vessels and a single bicarotid trunk, as opposed to the more common human aortic arch which splits into a single brachiocephalic trunk, the left common carotid 19 and the left subclavian arteries. 28,29 20 21 Similar to their ovine counterparts, little to no research has been done explicating the dimensional differences 22 in aortic root diameter between bovines and humans, elucidating the need for further research in this area.

Coronary anatomy 25
The suitability of porcine species usage as an animal model in coronary arterial disease is well established, 26 with multiple morphological studies highlighting that porcine coronary vasculature is closely resemblant to that 27 of man. 33 In pigs, both coronary arteries arise from the aortic sinuses below the supravalvular ridge, as is seen 28 in human species, with one study highlighting that all tested porcine models showed right coronary artery 29 (RCA) dominance (humans typically exhibit RCA dominance of anywhere between 75 to 85%, depending on 30 the chosen study analysed). 34 As with their human counterparts however, certain inter-species variants are 31 present, and should be kept in mind whilst choosing a porcine animal model. 34,35

33
With regards to the coronary arterial system, in contrast to their porcine and human analogues, ovine species 34 primarily have a left coronary type circulation; ergo, the majority of the myocardium receives its blood supply 35 via branches of the left coronary artery. 36 However, given that ovines do not possess an extensive coronary 36 collateral network, it may be still suitable to utilise their models for research. More specifically, although there exists considerable literature that is descriptive of specific aspects of ovine cardiac anatomy, little to no 38 comparative research has been performed to elucidate the differences between ovine and human heart 39 models, highlighting a significant literature gap in this area. 36

IJMS
The coronary vasculature of bovine species has also been studied and documented. In all examined animals, 1 the coronary ostia were located beneath the sinotubular junction, as seen within their human counterparts. 37

2
The dimensions of coronary ostia are listed in Table 3, but it is important note that ovines are one of the most 3 common veterinary species to exhibit coronary artery anomalies, with examples of such abnormalities 4 including coronary-to-pulmonary artery fistulae and anomalous origin of the left coronary artery from the 5 pulmonary trunk. Consequently, their usage as animal models to mimic the human coronary system merits 6 careful scrutiny before findings can be extrapolated. 38,39 7 8 Suitability for use as animal clinical models in Type-A aortic 9 dissection research 10 Having explored the anatomical differences between ovine, bovine and the porcine species, the anatomic 11 feasibility of using these as animal models to better our understanding of Type-A AD treatment options shall 12 now be explored.

14
As aforementioned, Type-A ADs involve the ascending aorta, making this aspect of the model's anatomy 15 significantly important. Bovine aortic anatomy is particularly unhelpful for this pathology therefore, given the 16 marked differences compared to their human counterparts, as elucidated previously. 28

21
Between the ovine and porcine species, it appears that each share some features with that of humans, whilst also exhibiting some differences that impact their usage as animal models. For instance, whilst ovines 23 maintain a uniform aortic diameter similar to their human counterparts, their short immobile aorta could pose a 24 challenge to graft repair within animal models. 27 Conversely, in spite of the larger aorta within pigs, the aortic 25 diameter being nearly a fifth lesser than that of their human counterparts could also affect reproducibility of 26 findings to the latter. Consequently, it is difficult to assess which of ovine or porcine models is better for 27 modelling Type-A ADs, at least from the ascending aortic anatomy point-of-view.

29
The aortic valvular anatomy also holds certain significance when choosing an appropriate animal model, 30 particularly with AD tears extending proximally into the aortic root. 41 As aforementioned, variations in leaflet 31 thickness are of importance, as the heavy pressure loads exerted during clinical usage with can affect the 32 structural stability of the animal model. Consequently, species with relatively thinner valvular commissures, 33 such as in ovines, must be handled with due care, and it is for this reason that porcine models are preferred to 34 their counterparts.

36
Finally, the coronary vasculature of the aforementioned animal models also has particular relevance to the 37 pathology of Type-A ADs, especially with tears arising in the aortic root, or even with any more distal tears