A most widely used conduits for revascularization of

Transcription

1 Effects of External Stenting on Wall Thickening in Arteriovenous Bypass Grafts Adonis G. Violaris, MRCP, Andrew C. Newby, PhD, and Gianni D. Angelini, FRCS Cardiac Surgery Department, University of Sheffield, Sheffield, and Cardiology Department, University of Wales College of Medicine, Clrdiff, United Kingdom Late occlusion of the saphenous vein graft appears to result in part from wall thickening as an adaptation to increased mean wall stress. Using an established pig model of arteriovenous bypass grafting, the effect of reducing wall stress with an external porous polytetrafluoroethylene stent was investigated. Segments of autologous saphenous vein were implanted by end-to-end anastomoses into both carotid arteries, with one graft supported by a stent 4 mm in diameter. Increases in graft wall dimensions were quantified 4 weeks later by computer-aided planimetry of transverse histological sections. The contribution of hyperplasia (i.e., cell proliferation) to the changes observed was further clarified by measurements of DNA concentration. All grafts showed an increase in external size, but this was restricted by stenting. All grafts also showed an increase in cross- sectional area of the tunica media and tunica intima that was only partly accounted for by an increase in DNA concentration, which indicated that both hyperplasia and hypertrophy had occurred. Stented grafts showed less enlargement of the media but greater enlargement of the intima. Overall wall size was therefore similar in stented and unstented grafts. Stented grafts showed less increase in DNA concentration than unstented grafts. In stented grafts, the residual luminal cross-sectional area was significantly less than in unstented grafts. The data show that external stenting reduces medial enlargement and hyperplasia but increases encroachment of the intima into the lumen. Because final luminal size is thought to be of paramount importance in maintaining long-term patency, external stenting is unlikely to be of benefit. (Ann Thorac Surg ) utologous saphenous vein continues to be one of the A most widely used conduits for revascularization of coronary and peripheral arteries despite its disappointing long-term patency rate [l-31. Late occlusion appears to be the result of both the intimal thickening caused by migration and proliferation of smooth muscle cells and the appearance of mature lipid-laden atherosclerotic plaques [3-51. Both these processes probably represent adaptive changes by the vein to its insertion into the arterial system [3-51. Factors potentially responsible for these changes may include the increased mean wall tension and the increased mean shear stress. If so, the use of an external support to limit the effect of wall tension might reduce the degree of wall thickening. To test this hypothesis, we used a previously described pig model of arteriovenous bypass grafting [6] to quantify the effects of a rigid external polytetrafluoroethylene support on the changes in intimal and medial size of the vein graft and in cell numbers reflected by DNA concentration. Material and Methods Surgical Procedures Studies were performed using 15 white race pigs initially weighing 20 to 25 kg. Premedication, induction and maintenance of anesthesia, and autologous saphenous Accepted for publication June 18, 1992 Address reprint requests to Mr Angelini, Department of Cardiac Surgery, University of Bristol, Bristol Royal Infirmary, Bristol BS2 BHW, United Kingdom. vein to carotid artery bypass grafting were performed as detailed previously [6]. Briefly, a longitudinal incision was made on the outer aspect of the hind leg to expose approximately 12 cm of the long saphenous vein. The vein was dissected free from surrounding tissue using a notouch technique [7], and all side branches were secured with a 6-0 Prolene ligature. Then the vein was removed, gently irrigated with a heparinized isosmotic sodium chloride solution (0.9 g/l), and stored in the same solution at room temperature for 20 to 40 minutes until needed. The common carotid artery was isolated through a longitudinal neck incision, and a 3- to 4-cm segment, isolated between vascular clamps, was excised. Both cut ends were beveled to approximately 45 degrees. A segment of reversed saphenous vein was similarly beveled and then anastomosed end-to-end to the carotid artery using a continuous 7-0 Prolene suture (Ethicon, Somerville, NJ). On the opposite side, after the proximal anastomosis was completed, the vein was passed through a polytetrafluoroethylene tube (Gore-Tex, W. L. Gore & Associates Inc, Elkton, MD) 4 mm in diameter and slightly longer than the graft, after which the distal anastomosis was performed. A single 6-0 Prolene suture was used to secure the polytetrafluoroethylene graft to the adventitial tissue of the two divided ends of the carotid artery. The neck and leg wounds were closed in layers, and the animals were allowed to recover. After 4 weeks, the animals were anesthetized as before, the neck was opened, and the graft was identified. The carotid artery was transected distal to the graft, and the by The Society of Thoracic Surgeons /93/! 6.00

2 668 VIOLARIS ET AL Ann Thorac Surg 1993:55: absence of blood flow was taken to indicate graft occlusion. Only patent grafts were used for analysis because graft occlusion was associated with extensive fibrotic changes that obscured the vein structure. Microscopic Methods The graft was removed and divided in two, and the larger segment was cannulated at one end with a syringe attached to a mercury manometer [8]. The other end of the graft also was cannulated and the cannula, was clamped. Fixative, consisting of a solution of 10% formalin in 0.1 mom sodium phosphate buffer, ph 7.3, was pumped into the lumen of the vessel for 5 to 10 minutes to achieve a pressure of approximately 100 mm Hg. The smaller segment was frozen in liquid nitrogen for estimation of DNA concentration. A segment of long saphenous vein from the other hind leg, ie, ungrafted vein, was obtained at the time of graft removal and was treated similarly. Fixed grafts and ungrafted segments of vein were further fixed, dehydrated, cleared, and embedded in paraffin wax with their axis perpendicular to the cutting plane. Sections were cut at four levels approximately 0.75 mm apart, mounted on glass slides, stained with Miller s elastic stain, and counterstained with van Gieson s stain. Luminal, intimal, and medial crosssectional areas were measured for three sections at each level by computer-aided planimetry using a Nikon microscope with a camera ludda, Hewlett-Packard computer (HP 85; Corvallis, OR), a digitizing tablet, and Graphic Information System quantitation program V3.0. The area enclosed by the endothelium defined the lumen. The area between the endothelium and the internal elastic lamina defined the intima. The area between the internal and external elastic laminae defined the media. Luminal encroachment was defined as the percentage of the area enclosed by the internal elastic lamina occupied by the intima. Mean values were then calculated for all sections from the same graft. Average intimal thickness was calculated from the cross-sectional area data using the following formulas and assuming that the sections consisted of circular profiles, which was a valid assumption because the tissues were fixed at normal perfusion pressure: (1) Luminal radius = vluminal area/.rr (2) Intimal thickness = v(intimal+ luminal area)/?r - luminal radius (3) Medial thickness = v(meciial+ intimal + luminal area)/.rr - luminal radius - intimal thickness DNA Concentration Frozen vessels were thawed, the adventitia was removed, and the remaining tissue was weighed and then homogenized at 4 C using a motor-driven Teflon glass homogenizer with 100 volumes of 0.3 mom perchloric acid until a smooth suspension was created. The DNA concentrations were then measured as detailed previously and were expressed as micrograms per milligram of wet weight [6]. The total DNA concentration (&mm graft length) was calculated as the product of DNA (&mg wet weight) multiplied by wall area (mm ), the assumption being that the tissue density was 1 mg/mm3. Echocardiographic Doppler Sonography A Sonicaid Vasoflo 111 continuous-wave Duplex ultrasound system (Oxford Sonicaid, Chichester, West Sussex, England) was used to produce simultaneously a highresolution image and the flow pattern within the grafts 10 to 20 minutes after completion of the surgical procedure, as previously described [6]. The diameters of the carotid artery, the in situ saphenous vein, and the grafts were measured at five points along their length, and the mean diameter was calculated. Statistical Analysis Values are shown as the mean 2 the standard error of the mean with the number of observations in parentheses. Unless otherwise stated, significance between groups was established with Student s t test using unpaired values. Results All operations were performed by the same surgeon (G.D.A.). One pig failed to survive until the grafts were recovered and hence was excluded from the analyses. Unstented grafts had a patency rate of 93% (n = 14) and stented grafts, 79% (n = 14) (p = not significant). Diameter Measurements and Flow Pattern Ultrasound measurements were made to determine the effect of stenting on the initial diameter of the vein grafts. The mean luminal diameter of the proximal and distal carotid artery at the level of grafting was mm and 3.0 k 0.1 mm, respectively (n = 5). Ungrafted in situ saphenous vein had a mean luminal diameter of mm (n = 5). Unstented grafts had a mean diameter of mm (n = 4) and stented grafts, mm (n = 4). Ultrasound measurements were also used to determine whether stenting affected flow patterns in the grafts. Velocity at peak systole and diastolic flow were comparable in the proximal carotid artery and in both unstented and stented grafts, implying similar flow rates (Fig 1). Flow inhomogeneity (related to turbulence) quantified by Doppler spectral broadening at peak systole was similar in the middle point of both unstented and stented grafts (54.1%? 0.5% and 53.8% 2 0.2%, respectively; n = 4). Changes in Luminal, Medial, and Zntimal Size and DNA Concentration In ungrafted vein, the intima consisted of a confluent monolayer of endothelium. The media, which was composed of an average of ten smooth muscle cell layers, was surrounded by well-defined internal and external elastic laminae (Fig 2A).

3 Ann Thorac Surg VIOLARIS ET AL 669 Table 1. Changes in Medial, lntimal, and Luminal Areas" Variable Vein Graft graft (n = 24) (n = 13) (n = 11) Medial area (mm') 1.1 f f 0.63b 2.25 f 0.33b,' Intima1 area (mm') b O.ab Total wall area 1.1? f 0.85b b (mm') Luminal area (mm') f 1.33b ' Luminal f 2.7b 55.3 f 5.Zb*' encroachment (%) a Data are shown as the mean f the standard error of the mean; the number of observations is in parentheses. Significance: p < 0.01 versus vein. ' Significance: p < versus unstented graft. Fig 1. Sonograms illustrating spectral flow pattern from (A) unstented and (B) stented grafts. In unstented and stented grafts recovered after 4 weeks, the internal elastic lamina remained unchanged, whereas the external elastic lamina became more separated (Figs 2B, 2C). A new intimal layer with both axially and circumferentially oriented smooth muscle cells was clearly present in both unstented and stented grafts (see Figs 2B, 2C). Medial and intimal cross-sectional areas were increased by grafting. Medial cross-sectional areas increased significantly more in unstented grafts (4.6-fold) compared with stented grafts (twofold) (Table 1). In contrast, the final intimal cross-sectional area was only about half as great in unstented as in stented grafts, although the difference did not reach significance (p = 0.1). As a result of these inverse changes, there was no significant difference in total wall area (intimal + medial cross-sectional areas) between unstented and stented grafts. Luminal areas increased approximately fivefold in unstented grafts compared with veins but remained virtually unchanged in stented grafts (see Table 1). As a result, luminal encroachment was significantly less in unstented than stented grafts (see Table 1). The DNA concentration per milligram of wet weight was lower in all grafts than in ungrafted vein (Table 2), as demonstrated previously [9]. This indicated that the density of viable cells had declined. The DNA concentration per milligram of wet weight was significantly greater in unstented than stented grafts (see Table 2), which indicated that the final density of viable cells was greater in unstented grafts. The calculated DNA concentration per Fig 2. Transverse sections of (A) ungrafted vein, (B) unstented graft, and (C) stented graft. (Miller's elastin stain and van Gieson's stain as counterstain; X75 before 24% reduction.) (A = adventitial tissue; EEL = external elastic lamina; IEL = internal elastic lamina; M = media; PTFE = polytetra/7uoroethylene.) B

4 670 VIOLARIS ET AL EXTERNAL STENTING OF VEIN GRAFI'S Ann Thorac Surg millimeter of length was sigtuficantly increased by grafting (see Table 2), which demonstrated that grafting caused an overall increase in vessel wall cell numbers. The DNA concentration per millimeter of length was significantly greater in unstented than stented grafts (see Table 2) as a result of the increased cell density. The results of the planimetry measurements were also used to calculate mean external radii and medial and intimal thicknesses. Consistent with the increase in graft size, unstented grafts showed a 2.4-fold increase in external radius (Table 3). This was much reduced in stented grafts (1.6-fold). Wall thickness (intimal + medial thicknesses) was increased 4.1-fold in unstented grafts and significantly more, 6.1-fold, in stented grafts (see Table 3). The difference between unstented and stented grafts was due to greater intimal thickening partly compensated by less medial thickening in the stented grafts (see Table 3). As a net result, luminal radius was significantly less in stented than unstented grafts (see Table 3). Comment Late saphenous vein graft failure remains a major problem in coronary and peripheral artery revascularization with up to 50% of vein grafts occluded by 10 years postoperatively [2, 3, 101. Previous studies have demonstrated that short-term graft patency may be improved by antiplatelet agents [3] or by optimal techniques of harvesting [6]. Late graft attrition, however, is influenced only marginally by antiplatelet agents or surgical technique [3, 9,101 and thus probably represents an intrinsic adaptation of the vein to its insertion in the arterial circulation. The higher pressure and flow velocities in the arterial circulation result in increases in both mean wall tension and mean shear stress. Either or both of these factors may cause smooth muscle cell migration into the intima and subsequent proliferation [4, 111. Surrounding the graft by a rigid external support has the theoretical advantage that wall tension will eventually be reduced or even eliminated. By limiting graft expansion, however, mean shear stress will be greater than in unstented grafts. The axial pressure pulse will be unaltered. The net effects of these differences are difficult to predict, and therefore the effect of stenting on the changes in wall dimensions was investigated empirically. Table 2. Changes in DNA Concentrations" Vein Graft Graft Variable (n = 24) (n = 13) (n = 11) Measured DNA 2.7 f f f O.Wb,' concentration (Fg/mg wet wt) Calculated DNA 1.97 f b 3.87 f 0.7ObPc concentration (pg/mm length) Data are shown as the mean +- the standard error of the mean; the number of observations is in parentheses. " Significance: p C 0.05 versus vein. ' Significance: p < 0.05 versus unstented graft. Table 3. Changes in Wall Dimensions" Vein Graft Graft Variable (n = 24) (n = 13) (n = 11) External radius 1.02 f f 0.09b 1.60 f 0.OT Wall thickness 0.14 f f 0.07b 0.85 f 0.09b,C Medial thickness 0.14 f f 0.05b 0.24 f 0.03' Intimal thickness f 0.04b 0.61 f O.lOb,' Luminal radius 0.98 f f 0.12b OT a Data are shown as the mean f the standard error of the mean; the number of observations is in parentheses. "Significance: p < 0.01 versus vein. ' Significance: p < 0.05 versus unstented graft. The results demonstrated that grafting led to an increase in vessel wall thickness accompanied by an increase in DNA concentration per millimeter of length. Thus hyperplasia of smooth muscle cells contributed to the increase in wall size. Nevertheless, the DNA concentration per milligram of wet weight fell, which implies that hypertrophy of cells and laying down of extracellular matrix also took place. Stenting with relatively rigid polytetrafluoroethylene had little net effect on wall thickening because lesser medial thickening was accompanied by greater intimal thickening. Stenting resulted in a reduction in DNA concentration per millimeter of graft length, which implies that less hyperplasia occurred and is consistent with the concept that wall tension is a determinant of hyperplasia. Stenting was also accompanied by a reduction in DNA concentration per milligram of wet weight, which implies that stented grafts had a relatively greater degree of hypertrophy or deposition of extracellular matrix. Owing to the restricted expansion of the stented grafts, the external radius of grafts was reduced by stenting. This together with a greater degree of intimal thickening resulted in greater encroachment into the vessel lumen and a reduction in residual luminal size. Stenting therefore increased predicted mean shear stress at the endothelial surface, although based on previous studies, this is associated with less rather than more intimal proliferation in both arteries [12] and arteriovenous bypass grafts [ll]. Alternative explanations for the greater intimal thickening we observed are that preventing graft expansion forced smooth muscle cells into the intima physically, or that interruption of regrowth of adventitial vessels may have caused greater hypertrophy, consistent with previous work [13]. Two other studies [14, 151 have examined the effect of external stenting on vein graft adaptation to the arterial circulation. Barras and colleagues [14] demonstrated that stenting decreased intimal thickness in a sheep model. The number of animals studied was small, however, and different animals were used for test and controls. In addition, the much thinner jugular vein was used, and the grafts were not pressure fixed. Perhaps for these

5 Ann Thorac Surg VIOLARIS ET AL 671 reasons, Barras and colleagues were unable to locate the internal elastic lamina and therefore measured intimal thickness as the distance from the subendothelial basal lamina to the beginning of the elastic external layer, which is equivalent to total wall thickness in our study. Kohler and colleagues [15] examined the effect of rigid external support on a rabbit model of arteriovenous bypass grafts. They also were unable to distinguish between the media and intima, but concluded that stenting did not reduce overall wall thickness. In agreement with our study, greater luminal encroachment occurred in the stented grafts, and this led to a smaller final luminal size. The study of Kohler and colleagues [15], however, was compromised by the fact that the proximal part of the graft was wrapped with polytetrafluoroethylene and the distal part of each graft was left to act as the corresponding control. The wrapping of the proximal part of the graft may have altered the local flow dynamics and may have influenced the degree of wall thickening in the control section. We confirmed that contralateral stented and unstented grafts had similar flow patterns, and so the differences between stented and unstented grafts are unlikely to result from differences in initial flow patterns. In conclusion, this study demonstrates that although external support leads to a reduction in smooth muscle hyperplasia, intimal thickening is promoted. This suggests that other factors such as constraining expansion of the vessel, greater shear stress, or interruption of the adventitia by the stent are also important determinants of wall thickening. The balance of factors is such that stenting of grafts leads to reduced luminal size. If directly applicable to the clinical situation, our results suggest that stenting is unlikely to increase long-term graft patency. Supported by grants from the British Heart Foundation and the National Heart Research Fund. References 1. Angelini GD, Bryan AJ, West RR, Newby AC, Breckenridge IM. Coronary artery bypass surgery: current practice in the United Kingdom. Thorax 1989;44: Lytle BW, Loop FD, Cosgrove DM, Ratliff NB, Easley K, Taylor PC. Long-term (5 to 12 years) serial studies of internal mammary artery and saphenous vein coronary bypass grafts. J Thorac Cardiovasc Surg 1985;89: Angelini GD, Newby AC. The future of saphenous vein as a coronary artery bypass conduit. Eur Heart J 1989;10:27? Zwolak RM, Adams MC, Clowes AW. Kinetics of vein graft hyperplasia: association with tangential stress. J Vasc Surg 1987;5: Boerboom LE, Bonchek LI, Kissebah AH, et al. Effect of surgical trauma on lipids in primate vein grafts: relation to plasma lipids. Circulation 1980;62(Suppl 1): Angelini GD, Bryan AJ, Williams HMJ, Morgan R, Newby AC. Distension promotes platelet and leukocyte adhesion and reduces short-term patency in pig arteriovenous bypass grafts. J Thorac Cardiovasc Surg 1990;99: Gottlob R. The preservation of venous endothelium by dissection without touching and by an atraumatic technique of vascular anastomosis. Minerva Chir 1977;32: Angelini GD, Passani SL, Breckenridge IM, Newby AC. Nature and pressure dependence of damage induced by distension of human saphenous vein coronary artery bypass grafts. Cardiovasc Res 1985;1932M. 9. Angelini GD, Bryan AJ, William HM, et al. Timecourse of medial and intimal thickening in pig arteriovenous bypass grafts. Relationship to endothelial injury and cholesterol accumulation. J Thorac Cardiovasc Surg 1992;103:109~ Campeau L, Enjalbert M, Lesperance J, Vaislic C, Grondin CM, Bourassa MG. Atherosclerosis and late closure of saphenous vein grafts: sequential angiographic studies 2 weeks, 1 year, 5 to 7 years and 10 to 12 years after surgery. Circulation 1983;68(S~ppl 2): Morinaga K, Okadome K, Kuroki M, Miyazaki T, Muto Y, Inokuchi K. Effect of wall shear stress on intimal thickening of arterially transplanted autogenous veins in dogs. J Vasc Surg 1985;2:43& Friedman MH, Hutchins GM, Bargeron CB, Deters OJ, Mark FF. Correlation between intimal thickness and fluid shear in human arteries. Atherosclerosis 1981;39: Karayannalos PE, Hostetler JR, Bono MG, et al. Late failure in vein grafts: mediating factors in subendothelial fibromuscular hyperplasia. Ann Surg 1978; Barras JA, Volant A, Leroy JP, et al. Constrictive perivenous mesh prosthesis for preservation of vein integrity. Experimental results and application for coronary bypass grafting. J Thorac Cardiovasc Surg 1986;92:33& Kohler TR, Kirkman TR, Clowes AW. The effect of rigid external support on vein graft adaptation to the arterial circulation. J Vasc Surg 1989;9: