Biorheology 39 (2002) 507–512 507IOS Press

The geometry of unstented and stented pigcommon carotid artery bypass grafts

Colin Caroa,∗, Jamie Jeremyc, Nick Watkinsa, Richard Bulbuliac, Gianni Angelinic,Frank Smithd, Song Wane, Anthony Yime, Spencer Sherwinb, Joaquim Peirób,Yannis Papaharilaoua,b, Brian Falzonb, Sergio Giordanab and Clydd GriffithsbaDepartment of Bioengineering, Imperial College, London, UKb Department of Aeronautics, Imperial College, London, UKc Bristol Heart Institute, University of Bristol, UKd Department of Surgery, Bristol Royal Infirmary, UKeDepartment of Surgery, Prince of Wales Hospital, Hong Kong

Abstract. The long-term success of arterial bypass grafting with autologous saphenous veins is limited by neointimal hyperpla-sia (NIH), which seemingly develops preferentially at sites where hydrodynamic wall shear is low. Placement of a loose-fitting,porous stent around end-to-end, or end-to-side, autologous saphenous vein grafts on the porcine common carotid artery hasbeen found significantly to reduce NIH, but the mechanism is unclear. In a preliminary study, we implanted autologous saphe-nous vein grafts bilaterally on the common carotid arteries of pigs, placing a stent around one graft and leaving the contralateralgraft unstented. At sacrifice 1 month post implantation, the grafts were pressure fixed in situ and resin casts were made. Un-stented graft geometry was highly irregular, with non-uniform dilatation, substantial axial lengthening, curvature, kinking, andpossible long-pitch helical distortion. In contrast, stented grafts showed no major dilatation, lengthening or curvature, but therewas commonly fine corrugation, occasional slight kinking or narrowing of segments, and possible long-pitch helical distortion.Axial growth of grafts against effectively tethered anastomoses could account for these changes. CFD studies are planned, using3D MR reconstructions, on the effects of graft geometry on the flow. Abnormality of the flow could favour the development ofvascular pathology, including NIH.

Keywords: Graft geometry, graft casts, graft elongation, graft buckling

1. Background and methods

Almost 50% of deaths in industrialised societies are attributable to atherosclerosis, the principal com-plications of which include myocardial infarction, stroke and peripheral vascular disease. Coronary arterybypass surgery (CABG) and infra-inguinal bypass graft surgery (IBBS) have revolutionised the treatmentof coronary artery disease and peripheral vascular disease, respectively, but the methodology is of greatimportance and has been the basis of extensive research.

Despite refinements in prosthetic materials, autologous saphenous vein remains the principal conduitfor CABG [1,17] and IBBS [14]. However, the long-term clinical success of arterial bypass grafting withautologous saphenous veins is limited by neointimal hyperplasia (NIH) and, eventually, by superim-posed atherosclerosis. Anti-platelet agents and minimization of vessel injury at the time of graft implan-tation [12,16] reduce the incidence of early thrombotic occlusion. Nevertheless, there is failure of over50% of coronary artery bypass grafts within 10 years [4] and of 25–40% of infra-inguinal grafts within

* Address for correspondence: Prof. C.G. Caro, Sir Leon Bagrit Centre, Department of Bioengineering, Imperial College,Exhibition Road, London SW7 2BX, UK. Tel.: +44 207 594 5180; Fax: +44 207 584 6897; E-mail: c.caro@ic.ac.uk.

0006-355X/02/$8.00 2002 – IOS Press. All rights reserved

508 C. Caro et al. / Geometry of bypass grafts

5 years [21] because of NIH. The biological/pathological changes accompanying NIH are the subject ofintensive study [14]. They include proliferation and migration of vascular smooth muscle cells acrossthe internal elastic lamina in response to endogenously generated growth factors, the deposition of extra-cellular matrix proteins and encroachment on the lumen and, ultimately, the formation of atheroscleroticplaque and graft stenosis.

The placement of a porous, non-restricting polyester DacronTM stent around saphenous vein–carotidartery interposition, or end-to-side, grafts in a porcine model significantly reduces NIH and total wallthickness [3,9,13,19]. Mechanisms considered to underlie this improvement include reduction of graftwall tangential (or hoop) stress by the stent [21]. However, the presence of a well-organized adventitialneovasculature in association with loose-fitting porous stents suggests that improvement of graft wallmass transport is a determinant of the reduction of NIH [9].

Recent investigations may shed further light on this question. The local flow field markedly influencesvascular biology [6,11] and, apparently, the development of vascular pathology [5,10,17]. Consistentwith these observations, NIH is reported to develop preferentially at sites at arterial bypass grafts wherehydrodynamic wall shear is low [2,20]. We have in a preliminary study constructed autologous saphenousvein grafts on both common carotid arteries in 12 healthy pigs (11 animals – end-to-end grafts, 1 animal– end-to-side grafts). In the ‘end-to-end’ animals, a porous non-restricting polyester DacronTM stent wasplaced around the graft on one carotid artery, while the contralateral carotid artery graft was unstented.In the ‘end-to-side’ animal, both grafts were unstented. At sacrifice, one month after surgery, the graftswere formalin fixedin situ at physiological transmural pressure, in order to preservein vivo geometryand, after their excision, resin casts were made (Biresin G49, Sika Chemie, GmBh) again at physiologicaltransmural pressure. The grafts were incised longitudinally to free the casts, the tissues being kept forhistological examination.

The geometry of the unstented graft casts, in particular, was highly irregular, involving non-uniformdilatation, curvature, kinking, and possible long pitch helical distortion (Fig. 1). The methods used tomeasure the 3D graft cast geometry were crude. Maximum and minimum graft cast diameters weremeasured at several stations and related to host artery cast diameter. The length of graft casts was deter-mined by running a fine solder wire axially along them and measuring its length when straightened. The

Fig. 1. Unstented graft cast showing dilatation, curvature, kinking, lengthening and helical distortion.

C. Caro et al. / Geometry of bypass grafts 509

grafts were taut at implantation, but neither graft length nor inter-anastomosis separation was measuredthen. As a means of assessing whether grafts lengthened following implantation, graft cast length wasrelated to inter-anastomosis separation, determined from the casts. The justification for this approachwas the assumption that tethering of the host artery proximally and distally would inhibit change ofinter-anastomosis separation.

2. Results

The unstented graft casts were irregularly highly curved, and showed in addition irregular dilatationand narrowing, kinking, and possible long pitch helical distortion. Away from sites of kinking, the graftswere approximately circular in cross-section, with diameters ranging from 0.6–1.0 cm to 0.3 cm. Nearsites of kinking, maximum diameters were 0.6–1.0 cm and minimum diameters were 0.3–0.4 cm. Thehost artery casts had diameters of 0.3–0.4 cm. Mean graft cast length was 4.33 cm, variance= 1.57,n = 13 and mean inter-anastomosis separation was 3.0 cm, variance 0.54. The ratio mean graft castlength/mean inter-anastomosis separation took a value of 1.45, variance= 0.09. The data for the twoend-to-side grafts are included in the above analysis. Their length ratios took values of 1.68 and 1.55,generally in line with those for the end-to-end grafts.

The stented graft casts tended to be finely corrugated, but generally showed no major curvature orkinking (Fig. 2). In several instances there was a suggestion of long pitch helical distortion. The diametersof the casts ranged from 0.4–0.5 cm. The host artery casts had diameters of 0.3–0.4 cm. In 3 graft casts,there were narrowed segments of length 0.5–1.0 cm, with diameters of about 0.2 cm (Fig. 3). Meangraft cast length was 2.91 cm, variance= 0.34, n = 11, and mean inter-anastomosis separation was2.62 cm, variance= 0.26. The ratio mean graft cast length/mean inter-anastomosis separation took avalue of 1.11, variance= 0.01.

The diameters of the unstented graft casts away from sites of kinking (range 0.6–1.0 cm) considerablyexceeded those of the stented graft casts away from narrowed segments (range 0.4–0.5 cm). The valueof the ratio mean graft cast length/mean inter-anastomosis separation was 1.45 for the unstented grafts,

Fig. 2. Stented graft cast showing slight lengthening, corrugation and possible helical distortion.

510 C. Caro et al. / Geometry of bypass grafts

Fig. 3. Stented graft cast showing a narrowed segment.

and 1.11 for the stented grafts. A 2-tailedt-test (two samples with unequal variance) showed the meansto be significantly different (p < 0.002).

3. Discussion

The geometry, particularly of the unstented graft casts, was highly complicated, and relatively crudemethods were used for its measurement. The complexity of the geometry also made evident the needfor additional measurements at the time of graft implantation. It seems possible nevertheless to derivesignificant conclusions from the work, given the validity of certain assumptions.

Graft diameter was not measured at implantation. However, if veins of similar diameter were used forthe unstented and stented grafts, it is permissible to conclude that the unstented grafts dilated substan-tially during the post-operative period, and that external stenting inhibited such dilatation. Graft lengthwas also not measured at implantation. However, if as is likely tethering of the host artery inhibitedchange of inter-anastomosis separation, it is permissible to conclude that the unstented grafts, in partic-ular, lengthened substantially during the post-operative period. The range of increase of length relativeto inter-anastomosis separation was for the unstented grafts 11–220% (mean 45%) and for the stentedgrafts 3–25% (mean 11%), implying that external stenting inhibited graft lengthening.

It is known that unstented vein arterial bypass grafts tend to dilate, but there does not appear to bewide recognition that they may also lengthen substantially. In an effort to clarify mechanisms, we acutelyelevated the transmural pressure of segments of human saphenous vein, excess to requirement for arterialbypass grafting, to 100 cm H2O. There was dilatation and an approximately 10% increase of length. Somesegments also showed slight kinking if their ends were clamped a fixed distance apart. Moreover, helicaldistortion was evident if the segments ran axially in a tube of diameter somewhat exceeding their own.

These results make it unlikely that passive mechanical extension accounted for the marked length-ening of the unstented grafts and suggest that there had been instead axial growth and re-modelling.External stenting inhibited dilatation, gross curvature, lengthening and buckling of the grafts. Neverthe-less, the graft casts showed fine corrugation and probable long-pitch helical distortion, consistent with

C. Caro et al. / Geometry of bypass grafts 511

Fig. 4. 3-Dimensional MR reconstruction of cast shown in Fig. 1.

axial growth constrained by effectively tethered anastomoses. The triggering of these various changes re-quires elucidation. Our conjecture is that it involves exposure of the vein grafts to arterial blood pressureand unaccustomed hydrodynamic wall shear, in the absence of adequate adventitial tethering.

There is accumulating evidence consistent with wall mass transport playing an important role in thedevelopment of NIH. As noted above, the process affects preferentially low wall shear regions [2,20] andaccompanying the marked reduction of NIH by loose-fitting porous external stents there was develop-ment of an abundant adventitial microvasculature [9].

We propose to use 3D MR surface reconstructions of unstented and stented grafts (Fig. 4) to investigatecomputationally the effect of luminal geometry on the flow. Pending such studies we speculate, on thebasis of the observed geometric distortion, that there was abnormality of the luminal flow field. If thatassumption is correct, loose-fitting porous external stents may additionally inhibit the development ofNIH by improving luminal flow, whether via effects of wall shear rate on blood:wall mass transport [7,8]or via effects of wall shear stress [6,11].

Acknowledgements

We acknowledge support from the British Heart Foundation, the Garfield Weston Foundation, theBUPA Foundation, Henry Smith’s Kensington Estate Charity, the Clothworkers’ Foundation, and theChinese University of Hong Kong.

References

[1] G.D. Angelini and A.C. Newby, The future of saphenous vein as a coronary artery bypass conduit,Eur. Heart J. 10 (1989),273–280.

[2] H.S. Bassiouny, S. White, S. Glagov, E. Choi, D.P. Giddens and C.K. Zarins, Anastomotic intimal hyperplasia: mechanicalinjury or flow induced,J. Vasc. Surg. 15 (1992), 708–717.

512 C. Caro et al. / Geometry of bypass grafts

[3] R.A. Bulbulia, F.C.T. Smith, C. Lloyd, G.D. Angelini and J.Y. Jeremy, Effect of external stents on neointima formation inporcine vein grafts implanted by end-to-side anastomosis,Br. J. Surg. 87 (2000), 11–12.

[4] L. Campeau, M. Enjalbert, J. Lesperance, C. Vaislic, C.M. Grondin and M.G. Bourassa, Atherosclerosis and late closureof aortocoronary saphenous vein grafts; sequential angiographic studies at 2 weeks, 1 year, 5–7 years and 10–12 yearsafter surgery,Circulation 68 (1983), 11-1–11-7.

[5] C.G. Caro, J.M. Fitz-Gerald and R.C. Schroter, Atheroma and arterial wall shear. Observation, correlation and proposalof a shear dependent mass transfer mechanism for atherogenesis,Proc. Roy. Soc. B. 177 (1971), 109–159.

[6] P.F. Davies, Flow-mediated endothelial mechanotransduction,Physiol. Rev. 75 (1995), 519–560.[7] R.O. Dull and P.F. Davies, Flow modulation of agonist (ATP)-response (Ca2+) coupling in vascular endothelial cells,Am.

J. Physiol. 261 (1991), H149–H154.[8] R.O. Dull, J.M. Tarbell and P.F. Davies, Mechanisms of flow-mediated signal transduction in endothelial cells: Kinetics

of ATP surface concentrations.J. Vasc. Res. 29 (1992), 410–419.[9] S.J. George, M.B. Izzat, P. Gadsdon, J.L. Johnson, A.P. Yim, S. Wan, A.C. Newby, G.D. Angelini and J.Y. Jeremy, Macro-

porosity is necessary for the reduction of neointimal and medial thickening by external stenting of porcine saphenous veinbypass grafts,Atherosclerosis 155 (2001), 329–336.

[10] D.P. Giddens, T.D. Tang and F. Loth, Fluid mechanics of arterial bifurcations, in:Biological Flows, M. Jaffrin andC.G. Caro, eds, Plenum Publishing Corporation, New York, 1995, pp. 51–68.

[11] M.A. Gimbrone, Vascular endothelium, hemodynamic forces and atherogenesis,Amer. J. Path. 155 (1999), 1–5.[12] D.H. Israel, P.C. Adams, B. Stein, J.H. Chesebro and V. Fuster, Anti-thrombotic therapy in the coronary vein graft patient,

Clin. Cardiol. 14 (1991), 283–295.[13] M.B. Izzat, D. Mehta, A.J. Bryan, B. Reeves, A.C. Newby and G.D. Angelini, The influence of external stent size on early

medial and neointimal thickening in a pig model of saphenous vein bypass grafting,Circulation 94 (1996), 1741–1745.[14] J.Y. Jeremy, D. Mehta, A.J. Bryan, D. Lewis and G.D. Angelini, Platelets and saphenous vein graft failure following

coronary bypass surgery,Platelets 8 (1997), 295–309.[15] J. Kunlin, Le traitement de l’arterite oblierante par la greffe veineuse,Arch. Mal. Coeur 42 (1949), 371–372.[16] F.W. LoGerfo, W.C. Quist, N.L. Cantelmo and C.C. Haudenschild, Integrity of vein grafts as a function of initial intimal

medial preservation,Circulation 68 (1983), 117–124.[17] Z. Mallat and A. Tedgui, Current perspective on the role of apoptosis in atherothrombotic disease,Circ. Res. 88 (2001),

998–1003.[18] D. Mehta, M.B. Izzat, A.J. Bryan and G.D. Angelini, Towards the prevention of vein graft failure,Int. J. Cardiol. 62

(1997), SS56–SS63.[19] D. Mehta, S.J. George, J.Y. Jeremy, M.B. Izzat, K.M. Southgate, A.J. Bryan, A.C. Newby and G.D. Angelini, External

stenting reduces long-term medial and neointimal thickening and platelet derived growth factor expression in a pig modelof arteriovenous bypass grafting,Nature Med. 4 (1998), 235–239.

[20] V.S. Sottiurai, J.S.T. Yao, R.C. Batson, S.L. Sue, R. Jones and Y.A. Nakamura, Distal anastomotic intimal hyperplasia:histopathologic character and biogenesis,Ann. Vasc. Surg. 3 (1989), 26–33.

[21] K. Varty, K.E. Allen, P.R.F. Bell and N.J.M. London, Infra-inguinal vein graft stenosis,Br. J. Surg. 80 (1993), 825–833.[22] R. Zwolak, M. Adams and A. Clowes, Kinetics of vein graft hyperplasia: association with tangential stress,J. Vasc. Surg.

5 (1987), 126–136.