Host A
Response
Comparison
Nicholas L.
to
Implanted
Between Mesh and Velour
Tilney, MD, Paul
J.
Boor, MD
\s=b\ Segments of open Dacron mesh grafts were subcutaneously implanted in rats and harvested for a period of up to 12 weeks after operation at serial intervals. The gross and histologic events of the host response to the external surface were compared to that of segments of low-porosity Dacron velour implanted in a similar fashion. Mature collagen, generously vascularized with new capillaries, was noted throughout the mesh within three to four weeks, while a tightly bonded inner fibrous layer had formed from the surrounding tissues. Major segments of velour floated in amorphous caseous material for up to five weeks. These pools of debris with their concomitant inflammatory response slowly resolved during a ten-week period. This prolonged healing may contribute to eventual graft closure by progressive fibro-
external surface and through the graft material itself have received less emphasis. In the present study, segments of Sparks mandril Da¬ cron grafts were implanted subcutaneously in rats and harvested at serial time intervals. The gross and histologie events noted were compared to those occurring around Dacron velour grafts of low porosity, implanted in a sim¬ ilar fashion. Such serial observations in animals may im¬ prove understanding of the biologic behavior of these syn¬ thetic materials as well as facilitating the development of new ones.
MATERIALS AND METHODS
sis and extrinsic contracture.
(Arch Surg 110:1469-1472, 1975)
synthetic prostheses available bypass Theplacement large-caliber, high-flow
or re¬
Autologous grafts, generated by implanting
plastic
for
of
Dacron Grafts
arteries have
been generally satisfactory, both in long-term stability and in patency.1 In smaller arteries with lesser flow (such as those in the lower extremities), fabric grafts may fail because of kinking with joint flexion, external compres¬ sion by fibrous tissue, or deterioration of the pseudoin¬ tima.2-6 The saphenous vein has been considerably more successful as an arterial conduit in the legs, although its use may be limited by absence or small diameter. a
mesh tube in the subcutaneous tissue (as first described by Sparks78) are intriguing and successful alternatives. Such grafts have been used successfully for femoropopliteal by¬ pass and for access sites for hemodialysis.9 The techniques of implantation, harvesting, and anastomosis of such con¬ duits have been well described. However, the dynamics of cellular infiltration and the formation of collagen from the Accepted for publication March 4, 1975. From the departments of surgery (Dr Tilney) and pathology (Dr Boor), Peter Bent Brigham Hospital and Harvard Medical School, Boston. Reprint requests to Department of Surgery, Peter Bent Brigham Hospital, 721 Huntington Ave, Boston, MA 02115 (Dr Tilney).
Outbred male rats weighing between 350 and 400 gm were used in this study. The rats came from a closed laboratory population. Two types of vascular prostheses were tested: (1) 6-mm Sparks mandril Dacron mesh and (2) Dacron velour. The grafts were 3 cm in length, containing dimethicone silicone elastomer mandrils that were held in place by silk ties at each end. Comparable segments of each graft were studied in two populations of rats. The material of the Sparks graft is comprised of two layers of siliconized, knitted two-ply, hard-twist Dacron thread. The outer layer, knitted on a 22 needle head, provides strength; the inner layer, with larger interstices knitted on a 16 needle head, acts as a spacer for ingrowing tissues between the outer layer and the sili¬ cone rod.7 The complete porosity of the graft allows both the free ingrowth of host cells and the easy ingress of capillaries for perfu¬ sion of the matured collagenous wall. The Dacron velour was cho¬ sen as a low-porosity material. The multiple open fibrillar loops that project above the body of the yarn, especially on the outer surface, are designed to trap fibrin and cells, and to insure binding to surrounding tissues. The graft interstices are crisscrossed and subdivided by fibrils.'" This contrasts to the small but open inter¬ stices of smooth prostheses, and it may reduce bleeding at time of
operation.
The animals were anesthetized with ether, and their backs and flanks were clipped free of hair. A small incision through the skin and panniculus carnosus muscle was made in each subcostal space. A subcutaneous pocket was created by blunt dissection, and the graft was implanted. The skin was closed with 4-0 silk, and the in¬ cision was sprayed with Neosporin aerosol ( a mixture of poly-
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
myxin B sulfate, neomycin sulfate, and gramicidin). All animals given food and water ad libitum. The mesh grafts were harvested in duplicate from 12 rats at 2, 4, 8, and 10 days, and at 3, 4, 6, 10, and 12 weeks after placement. The velour grafts were gathered in duplicate from six rats at weekly intervals from two to ten weeks after placement. An el¬ liptical skin incision was made, and the skin, graft, and surround¬ ing tissues were removed en bloc. The vascularity of the adjacent tissues was assessed grossly. The tissue specimen was fixed in 10% buffered formaldehyde solution, and the silicone elastomer cores were removed from the grafts. After embedding in paraf¬ fin, multiple sections were stained with hemotoxylin-eosin, triwere
chrome, and with fibrin and elastic tissue stains. The sections later examined by light microscopy. Selected sections also stained with PAS, Gram, and alizarin red S. were
were
RESULTS
Sparks
Mandril
At two and four days after implantation, the graft lay free from surrounding tissues that were smooth, thick¬ ened, and without obvious inflammatory response. A small amount of clear fluid was present, while a thin and white gelatinous material, easily separable from the areolar sheath, infiltrated the mesh. Graft thickness, from the sili¬ cone elastomer rod to the adjacent subcutaneous fat, mea¬ sured 300a to 400µ at day 4. At 48 hours, the interstices were filled with fibrin and loculated serum. By the fourth day, an acute inflammatory response was present around the graft. Polymorphonuclear leukocytes had invaded the mesh, and a few lymphocytes and plasma cells were noted. The small vessels in the surrounding subcutaneous tissue were congested and showed swelling of the endothelial cells. Grafts removed eight and ten days after implantation adhered firmly to the tissue sheath. There was abundant neovasculature at the external surface, and the vessels to¬ tally encircled the graft segment in several places along its length. The adjacent areolar tissues were less thick than they were at day 4. Microscopically, marked, regu¬ larly distributed fibroblastic proliferation was noted throughout the mesh (Fig 1). Scanty mature collagen was present, as was shown by trichrome staining. The striking proliferation of small vessels around and through the mesh interstices was accompanied by a minimal acute and chronic inflammatory response. The cellular infiltrate in the surrounding tissues had begun to resolve. The gross and histologie pictures from weeks 3 to 12 were relatively unchanged. The surrounding tissues were not thickened, although the thickness of the graft from the silicone elastomer core to the adjacent fat had in¬ creased, and measured 400µ to 500µ. Considerable amounts of vascularized fibrous tissue were present throughout the mesh. Inflammation and fibroblastic proliferation were no longer evident in the graft or in the surrounding tissues. A rather consistent inner fibrous layer generally mea¬ sured 30µ to 80µ from the silicone elastomer core to the first strands of Dacron. However, in several places, the mesh was bare on the luminal side after removal of the mandril (Fig 2). No newly formed elastic tissue was ob¬ served in any graft, although collagen was abundant.
Dacron Velour
Velour grafts implanted with silicone elastomer cores demonstrated remarkable gross and microscopic differ¬ ences from the mesh grafts. At weeks 2, 3, and 5, portions of the velour grafts floated in abundant amorphous, fri¬ able, and caseous debris. This predominantly occurred between the silicone elastomer core and the graft mate¬ rial, sometimes separating these structures by 5 mm. The caseous material was frequently distributed outside the graft, thus preventing close proximity of the surrounding tissues and the graft wall (Fig 3). These pools of debris were composed of necrotic cells, fi¬ brin, and precipitated calcium, as demonstrated by aliza¬ rin red S stain. Although no organisms were demonstra¬ ble by Gram stain or PAS, the pools were surrounded by an intense and acute inflammatory response, with many polymorphonuclear leukocytes and foamy macrophages. In the areas uninvolved with these changes, small vessels had proliferated within the mesh at week 2. In addition, gran¬ ulation tissue had surrounded the graft with mononuclear cells, and macrophages were scattered diffusely through¬ out. Abundant mature collagen was noted around and within the graft material at week 3, although little subse¬ quent increase in collagen thickness occurred later. The lakes of amorphous material surrounding and infiltrating the graft resolved slowly, and at week 5, many granulocytes were still present among copious fibrin and cellu¬ lar debris. These sites persisted for as long as ten weeks, although they gradually diminished in size. Healing was greatly prolonged, as indicated by the complete absence of mature collagen in portions of the graft adjacent to debris and acute inflammation. Elastic tissue was not found. COMMENT
Although this study was performed in rats, the compari¬ between loose mesh and velour suggests an important cause of late graft failure in a clinical setting. The mesh graft healed rapidly, and well-vascularized, mature colla¬ gen developed, along with the formation of a smooth and son
adherent fibrous intima. This reaction contrasted with the unexpected delayed healing of the velour graft, which was associated with consideration deposition of necrotic debris and calcium. Although cellular ingrowth and host re¬ sponses may vary substantially between species, a similar necrotizing reaction has been described around prostheses implanted in humans, especially those prostheses of low porosity.1013 The intensity and prolonged nature of the in¬ flammatory response associated with these large necrotic areas may contribute to eventual occlusion of such prostheses by intense circumferential fibrosis of their outer layer, progressive contracture of surrounding tissues, and critical reduction in luminal size. The external velour sur¬ face was originally designed to encourage adherence to fi¬ brous tissue by decreasing the separation of graft from tissue capsule by seroma formation.1" We could not sub¬ stantiate these findings in this study. The formation of an autologous fibrocollagenous arte¬ rial conduit on a strut of loosely woven fiber is appealing,
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
Fig 1.—Dacron mesh Sparks mandril graft, removed ten days after implantation. Fibroblastic and capillary proliferation (arrow) is marked within loose weave. Small amounts of collagen are de¬ monstrable by special stain at this time (modified Verhoeff-Van Gieson elastic tissue stain, 100).
Fig 2.—Sparks mandril Dacron graft (mandril removed) is es¬ sentially mature three weeks after implantation in rat. Inner fi¬ brous layer is regular, although there is occasional apposition of Dacron fiber to mandril. Arrow indicates mandril (hematoxylineosin, x40).
especially in the extremities, since long-term patency for synthetic prostheses is relatively poor, and adequate sa¬ phenous veins are not always available. The open Dacron mesh provides a strong, flexible wall that allows free in¬
Fig 3.—Graft removed after ten weeks. Lower-power view shows "pool" of dark-staining necrotic debris between subcuta¬ neous tissues of host (SC) and Dacron velour material (arrow)
growth of vascularized fibrocollagenous tissue during the six-week period of maturation. The tearing of such tubes when they are unsupported by synthetic mesh has been noted,- owing to a lack of collagen fiber interlacing in the neoadventitia. The extensive neovasculature, similar to vasa vasorum, allows adequate perfusion and nutrition of the arterial wall, while a silicone elastomer mandril en¬ sures the formation of a thin and smooth, yet complete fi¬ brous intima. Unlike the pseudointima that inevitably forms on other types of plastic arterial prostheses, the fi¬ brous intima of the autologous die-grown graft binds to the prosthesis as an integral part. Although the growth of elastic fibers surrounding silicone elastomer implants has also been described,1' formation of this tissue around the grafts was not noted in these experiments. A pseudointima develops rapidly in functioning arterial grafts and remains a thin, translucent, and smooth fibrin layer.' An important reason for late closure of such grafts is the loose bonding of pseudointima to graft surface, with luminal occlusion occasionally occurring after disruption of this layer. While no such information is available in this model, the delayed and irregular healing of the Dacron ve¬ lour, with prolonged absence of mature collagen in the in¬ terstices, suggests that bonding may be late and may in¬ vite intimai disruption. This contrasts with the very tightly bonded fibrous intima of the Sparks graft, which forms an integral part of the entire wall. The formation of a pseudointima within the lumina of fabrics implanted into the blood stream has previously been described."'"-"·'·'' Following the layering of protein on the bare synthetic fibers, fibrin and platelets are selec¬ tively deposited on the surface. Layers of developing thrombus are rapidly washed away by the blood flow in
(hematoxylin-eosin,
25).
but they may collect on the walls of vessels with lower flow. Leukocytes enter this thrombus layer and migrate more deeply toward and into the prosthetic wall, where they allegedly transform into fibroblasts.'-'"'"' These cells may organize and replace the thrombus in some areas of the graft, but may be absent in other areas. Their conjunction with cells growing in from the outside of the prosthesis may help to anchor the pseudointima. Capillaries that penetrate the fabric interstices may in¬ crease cell traffic and hasten maturation, although fewer capillaries can form with the tight weave of the prosthetic material. In some species, such as the baboon, capillaries open directly into the vascular lumen and provide a direct source of endothelial cells." The irregularity of this process has been noted by several authors.112 Another well-known hazard of the prosthetic graft is its
large vessels,
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
vulnerability
to infection.
Theoretically, complete
envel¬
opment and ingrowth of the graft by dense fibrous tissue may afford an effective barrier to bacteria. This concept is supported in part by recently reported clinical experience,
in which the rate of infection of Sparks mandrils used for chronic hemodialysis was quite minimal." This hypothesis is subject to testing through the use of standard doses of contaminating bacteria. The above-mentioned studies do not completely imitate the clinical setting, since the grafts were subcutaneous
implants rather than arterial conduits. Wesolowski and co-workers had detailed histologie changes in some grafts implanted in dogs, although they did not include studies of the Sparks mandril.13 They were able to demonstrate that, of all commonly used laboratory species, the host re¬ sponses of the dog are most similar to those of the human. The observations in this study on the response of the rat to a loose-mesh prosthesis may encourage the concept of the autologous fibrocollagenous conduit for more exten¬ sive experimental and clinical use.
References 1. DeBakey ME, Jordan GL Jr, Abbott JP, et al: The fate of Dacron vascular grafts. Arch Surg 89:757-782, 1964. 2. Parsonnet V, Alpert J, Brief DK: Autogenous polyprophylene-supported collagen tubes for long-term arterial replacement. Surgery 70:935\x=req-\ 939, 1971. 3. Hall KV: The arterial homograft used as "bypass" in patients with femoro-popliteal arteriosclerotic obstruction: A follow-up study. Acta Chir Scand 127:353-366, 1964. 4. Wesolowski SA, Fries CC, Henniger G, et al: Factors contributing to long-term failures in human vascular prosthetic grafts. J Cardiovasc Surg 5:544-567, 1964. 5. Kouchoukos NT, Levy JF, Balfour JF, et al: Operative therapy for femoral-popliteal arterial occlusive disease: A comparison of therapeutic methods. Circulation 35(suppl 1):174-182, 1967. 6. Warren R, McCoombs HL: Morphologic studies on plastic arterial prosthesis in humans. Ann Surg 161:73-82, 1965. 7. Sparks CH: Die-grown reinforced arterial grafts: Observations on long-term animal grafts and clinical experience. Ann Surg 172:787-792, 1970. 8. Sparks CH: Silicone mandril method of femoropopliteal artery bypass.
Am J Surg 124:244-249, 1972. 9. Morgan AP, Lazarus JM: Vascular access for dialysis: Techniques and newer methods. Am J Surg, to be published. 10. Sauvage LR, Berger K, Wood SJ, et al: An external velour surface for porous arterial prosthesis. Surgery 70:940-953, 1971. 11. Brais MP, Braunwald NS: Tissue acceptance of materials implanted within the circulatory system. Arch Surg 109:351-358, 1974. 12. Ghidoni JJ, Liotta D, Hall CW, et al: Healing of pseudointimas in velour-lined, impermeable arterial prosthesis. Am J Pathol 53:375-390, 1968. 13. Wesolowski SA, Fries CC, Karlson KE, et al: Porosity: Primary determinant of ultimate fate of synthetic vascular grafts. Surgery 50:91-96, 1961. 14. Hern\l=a'\ndez-Ja\l=u'\reguiP, Esperanza-Garci\l=a'\C, Gonz\l=a'\lez-AnguloA: Morphology of the connective tissue grown in response to implanted silicone rubber. A light and microscopic study. Surgery 75:631-637, 1974. 15. Florey HW, Greer SJ, Kiser J, et al: The development of the pseudointima lining fabric grafts of the aorta. Br J Exp Pathol 43:655-660, 1962. 16. Gillman T, Wright LJ: Autoradiographic evidence suggesting in vivo transformation of some blood mononuclears in repair and fibrosis. Nature 209:1086-1090, 1966.
results with
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
Response
Comparison
Nicholas L.
to
Implanted
Between Mesh and Velour
Tilney, MD, Paul
J.
Boor, MD
\s=b\ Segments of open Dacron mesh grafts were subcutaneously implanted in rats and harvested for a period of up to 12 weeks after operation at serial intervals. The gross and histologic events of the host response to the external surface were compared to that of segments of low-porosity Dacron velour implanted in a similar fashion. Mature collagen, generously vascularized with new capillaries, was noted throughout the mesh within three to four weeks, while a tightly bonded inner fibrous layer had formed from the surrounding tissues. Major segments of velour floated in amorphous caseous material for up to five weeks. These pools of debris with their concomitant inflammatory response slowly resolved during a ten-week period. This prolonged healing may contribute to eventual graft closure by progressive fibro-
external surface and through the graft material itself have received less emphasis. In the present study, segments of Sparks mandril Da¬ cron grafts were implanted subcutaneously in rats and harvested at serial time intervals. The gross and histologie events noted were compared to those occurring around Dacron velour grafts of low porosity, implanted in a sim¬ ilar fashion. Such serial observations in animals may im¬ prove understanding of the biologic behavior of these syn¬ thetic materials as well as facilitating the development of new ones.
MATERIALS AND METHODS
sis and extrinsic contracture.
(Arch Surg 110:1469-1472, 1975)
synthetic prostheses available bypass Theplacement large-caliber, high-flow
or re¬
Autologous grafts, generated by implanting
plastic
for
of
Dacron Grafts
arteries have
been generally satisfactory, both in long-term stability and in patency.1 In smaller arteries with lesser flow (such as those in the lower extremities), fabric grafts may fail because of kinking with joint flexion, external compres¬ sion by fibrous tissue, or deterioration of the pseudoin¬ tima.2-6 The saphenous vein has been considerably more successful as an arterial conduit in the legs, although its use may be limited by absence or small diameter. a
mesh tube in the subcutaneous tissue (as first described by Sparks78) are intriguing and successful alternatives. Such grafts have been used successfully for femoropopliteal by¬ pass and for access sites for hemodialysis.9 The techniques of implantation, harvesting, and anastomosis of such con¬ duits have been well described. However, the dynamics of cellular infiltration and the formation of collagen from the Accepted for publication March 4, 1975. From the departments of surgery (Dr Tilney) and pathology (Dr Boor), Peter Bent Brigham Hospital and Harvard Medical School, Boston. Reprint requests to Department of Surgery, Peter Bent Brigham Hospital, 721 Huntington Ave, Boston, MA 02115 (Dr Tilney).
Outbred male rats weighing between 350 and 400 gm were used in this study. The rats came from a closed laboratory population. Two types of vascular prostheses were tested: (1) 6-mm Sparks mandril Dacron mesh and (2) Dacron velour. The grafts were 3 cm in length, containing dimethicone silicone elastomer mandrils that were held in place by silk ties at each end. Comparable segments of each graft were studied in two populations of rats. The material of the Sparks graft is comprised of two layers of siliconized, knitted two-ply, hard-twist Dacron thread. The outer layer, knitted on a 22 needle head, provides strength; the inner layer, with larger interstices knitted on a 16 needle head, acts as a spacer for ingrowing tissues between the outer layer and the sili¬ cone rod.7 The complete porosity of the graft allows both the free ingrowth of host cells and the easy ingress of capillaries for perfu¬ sion of the matured collagenous wall. The Dacron velour was cho¬ sen as a low-porosity material. The multiple open fibrillar loops that project above the body of the yarn, especially on the outer surface, are designed to trap fibrin and cells, and to insure binding to surrounding tissues. The graft interstices are crisscrossed and subdivided by fibrils.'" This contrasts to the small but open inter¬ stices of smooth prostheses, and it may reduce bleeding at time of
operation.
The animals were anesthetized with ether, and their backs and flanks were clipped free of hair. A small incision through the skin and panniculus carnosus muscle was made in each subcostal space. A subcutaneous pocket was created by blunt dissection, and the graft was implanted. The skin was closed with 4-0 silk, and the in¬ cision was sprayed with Neosporin aerosol ( a mixture of poly-
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
myxin B sulfate, neomycin sulfate, and gramicidin). All animals given food and water ad libitum. The mesh grafts were harvested in duplicate from 12 rats at 2, 4, 8, and 10 days, and at 3, 4, 6, 10, and 12 weeks after placement. The velour grafts were gathered in duplicate from six rats at weekly intervals from two to ten weeks after placement. An el¬ liptical skin incision was made, and the skin, graft, and surround¬ ing tissues were removed en bloc. The vascularity of the adjacent tissues was assessed grossly. The tissue specimen was fixed in 10% buffered formaldehyde solution, and the silicone elastomer cores were removed from the grafts. After embedding in paraf¬ fin, multiple sections were stained with hemotoxylin-eosin, triwere
chrome, and with fibrin and elastic tissue stains. The sections later examined by light microscopy. Selected sections also stained with PAS, Gram, and alizarin red S. were
were
RESULTS
Sparks
Mandril
At two and four days after implantation, the graft lay free from surrounding tissues that were smooth, thick¬ ened, and without obvious inflammatory response. A small amount of clear fluid was present, while a thin and white gelatinous material, easily separable from the areolar sheath, infiltrated the mesh. Graft thickness, from the sili¬ cone elastomer rod to the adjacent subcutaneous fat, mea¬ sured 300a to 400µ at day 4. At 48 hours, the interstices were filled with fibrin and loculated serum. By the fourth day, an acute inflammatory response was present around the graft. Polymorphonuclear leukocytes had invaded the mesh, and a few lymphocytes and plasma cells were noted. The small vessels in the surrounding subcutaneous tissue were congested and showed swelling of the endothelial cells. Grafts removed eight and ten days after implantation adhered firmly to the tissue sheath. There was abundant neovasculature at the external surface, and the vessels to¬ tally encircled the graft segment in several places along its length. The adjacent areolar tissues were less thick than they were at day 4. Microscopically, marked, regu¬ larly distributed fibroblastic proliferation was noted throughout the mesh (Fig 1). Scanty mature collagen was present, as was shown by trichrome staining. The striking proliferation of small vessels around and through the mesh interstices was accompanied by a minimal acute and chronic inflammatory response. The cellular infiltrate in the surrounding tissues had begun to resolve. The gross and histologie pictures from weeks 3 to 12 were relatively unchanged. The surrounding tissues were not thickened, although the thickness of the graft from the silicone elastomer core to the adjacent fat had in¬ creased, and measured 400µ to 500µ. Considerable amounts of vascularized fibrous tissue were present throughout the mesh. Inflammation and fibroblastic proliferation were no longer evident in the graft or in the surrounding tissues. A rather consistent inner fibrous layer generally mea¬ sured 30µ to 80µ from the silicone elastomer core to the first strands of Dacron. However, in several places, the mesh was bare on the luminal side after removal of the mandril (Fig 2). No newly formed elastic tissue was ob¬ served in any graft, although collagen was abundant.
Dacron Velour
Velour grafts implanted with silicone elastomer cores demonstrated remarkable gross and microscopic differ¬ ences from the mesh grafts. At weeks 2, 3, and 5, portions of the velour grafts floated in abundant amorphous, fri¬ able, and caseous debris. This predominantly occurred between the silicone elastomer core and the graft mate¬ rial, sometimes separating these structures by 5 mm. The caseous material was frequently distributed outside the graft, thus preventing close proximity of the surrounding tissues and the graft wall (Fig 3). These pools of debris were composed of necrotic cells, fi¬ brin, and precipitated calcium, as demonstrated by aliza¬ rin red S stain. Although no organisms were demonstra¬ ble by Gram stain or PAS, the pools were surrounded by an intense and acute inflammatory response, with many polymorphonuclear leukocytes and foamy macrophages. In the areas uninvolved with these changes, small vessels had proliferated within the mesh at week 2. In addition, gran¬ ulation tissue had surrounded the graft with mononuclear cells, and macrophages were scattered diffusely through¬ out. Abundant mature collagen was noted around and within the graft material at week 3, although little subse¬ quent increase in collagen thickness occurred later. The lakes of amorphous material surrounding and infiltrating the graft resolved slowly, and at week 5, many granulocytes were still present among copious fibrin and cellu¬ lar debris. These sites persisted for as long as ten weeks, although they gradually diminished in size. Healing was greatly prolonged, as indicated by the complete absence of mature collagen in portions of the graft adjacent to debris and acute inflammation. Elastic tissue was not found. COMMENT
Although this study was performed in rats, the compari¬ between loose mesh and velour suggests an important cause of late graft failure in a clinical setting. The mesh graft healed rapidly, and well-vascularized, mature colla¬ gen developed, along with the formation of a smooth and son
adherent fibrous intima. This reaction contrasted with the unexpected delayed healing of the velour graft, which was associated with consideration deposition of necrotic debris and calcium. Although cellular ingrowth and host re¬ sponses may vary substantially between species, a similar necrotizing reaction has been described around prostheses implanted in humans, especially those prostheses of low porosity.1013 The intensity and prolonged nature of the in¬ flammatory response associated with these large necrotic areas may contribute to eventual occlusion of such prostheses by intense circumferential fibrosis of their outer layer, progressive contracture of surrounding tissues, and critical reduction in luminal size. The external velour sur¬ face was originally designed to encourage adherence to fi¬ brous tissue by decreasing the separation of graft from tissue capsule by seroma formation.1" We could not sub¬ stantiate these findings in this study. The formation of an autologous fibrocollagenous arte¬ rial conduit on a strut of loosely woven fiber is appealing,
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
Fig 1.—Dacron mesh Sparks mandril graft, removed ten days after implantation. Fibroblastic and capillary proliferation (arrow) is marked within loose weave. Small amounts of collagen are de¬ monstrable by special stain at this time (modified Verhoeff-Van Gieson elastic tissue stain, 100).
Fig 2.—Sparks mandril Dacron graft (mandril removed) is es¬ sentially mature three weeks after implantation in rat. Inner fi¬ brous layer is regular, although there is occasional apposition of Dacron fiber to mandril. Arrow indicates mandril (hematoxylineosin, x40).
especially in the extremities, since long-term patency for synthetic prostheses is relatively poor, and adequate sa¬ phenous veins are not always available. The open Dacron mesh provides a strong, flexible wall that allows free in¬
Fig 3.—Graft removed after ten weeks. Lower-power view shows "pool" of dark-staining necrotic debris between subcuta¬ neous tissues of host (SC) and Dacron velour material (arrow)
growth of vascularized fibrocollagenous tissue during the six-week period of maturation. The tearing of such tubes when they are unsupported by synthetic mesh has been noted,- owing to a lack of collagen fiber interlacing in the neoadventitia. The extensive neovasculature, similar to vasa vasorum, allows adequate perfusion and nutrition of the arterial wall, while a silicone elastomer mandril en¬ sures the formation of a thin and smooth, yet complete fi¬ brous intima. Unlike the pseudointima that inevitably forms on other types of plastic arterial prostheses, the fi¬ brous intima of the autologous die-grown graft binds to the prosthesis as an integral part. Although the growth of elastic fibers surrounding silicone elastomer implants has also been described,1' formation of this tissue around the grafts was not noted in these experiments. A pseudointima develops rapidly in functioning arterial grafts and remains a thin, translucent, and smooth fibrin layer.' An important reason for late closure of such grafts is the loose bonding of pseudointima to graft surface, with luminal occlusion occasionally occurring after disruption of this layer. While no such information is available in this model, the delayed and irregular healing of the Dacron ve¬ lour, with prolonged absence of mature collagen in the in¬ terstices, suggests that bonding may be late and may in¬ vite intimai disruption. This contrasts with the very tightly bonded fibrous intima of the Sparks graft, which forms an integral part of the entire wall. The formation of a pseudointima within the lumina of fabrics implanted into the blood stream has previously been described."'"-"·'·'' Following the layering of protein on the bare synthetic fibers, fibrin and platelets are selec¬ tively deposited on the surface. Layers of developing thrombus are rapidly washed away by the blood flow in
(hematoxylin-eosin,
25).
but they may collect on the walls of vessels with lower flow. Leukocytes enter this thrombus layer and migrate more deeply toward and into the prosthetic wall, where they allegedly transform into fibroblasts.'-'"'"' These cells may organize and replace the thrombus in some areas of the graft, but may be absent in other areas. Their conjunction with cells growing in from the outside of the prosthesis may help to anchor the pseudointima. Capillaries that penetrate the fabric interstices may in¬ crease cell traffic and hasten maturation, although fewer capillaries can form with the tight weave of the prosthetic material. In some species, such as the baboon, capillaries open directly into the vascular lumen and provide a direct source of endothelial cells." The irregularity of this process has been noted by several authors.112 Another well-known hazard of the prosthetic graft is its
large vessels,
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
vulnerability
to infection.
Theoretically, complete
envel¬
opment and ingrowth of the graft by dense fibrous tissue may afford an effective barrier to bacteria. This concept is supported in part by recently reported clinical experience,
in which the rate of infection of Sparks mandrils used for chronic hemodialysis was quite minimal." This hypothesis is subject to testing through the use of standard doses of contaminating bacteria. The above-mentioned studies do not completely imitate the clinical setting, since the grafts were subcutaneous
implants rather than arterial conduits. Wesolowski and co-workers had detailed histologie changes in some grafts implanted in dogs, although they did not include studies of the Sparks mandril.13 They were able to demonstrate that, of all commonly used laboratory species, the host re¬ sponses of the dog are most similar to those of the human. The observations in this study on the response of the rat to a loose-mesh prosthesis may encourage the concept of the autologous fibrocollagenous conduit for more exten¬ sive experimental and clinical use.
References 1. DeBakey ME, Jordan GL Jr, Abbott JP, et al: The fate of Dacron vascular grafts. Arch Surg 89:757-782, 1964. 2. Parsonnet V, Alpert J, Brief DK: Autogenous polyprophylene-supported collagen tubes for long-term arterial replacement. Surgery 70:935\x=req-\ 939, 1971. 3. Hall KV: The arterial homograft used as "bypass" in patients with femoro-popliteal arteriosclerotic obstruction: A follow-up study. Acta Chir Scand 127:353-366, 1964. 4. Wesolowski SA, Fries CC, Henniger G, et al: Factors contributing to long-term failures in human vascular prosthetic grafts. J Cardiovasc Surg 5:544-567, 1964. 5. Kouchoukos NT, Levy JF, Balfour JF, et al: Operative therapy for femoral-popliteal arterial occlusive disease: A comparison of therapeutic methods. Circulation 35(suppl 1):174-182, 1967. 6. Warren R, McCoombs HL: Morphologic studies on plastic arterial prosthesis in humans. Ann Surg 161:73-82, 1965. 7. Sparks CH: Die-grown reinforced arterial grafts: Observations on long-term animal grafts and clinical experience. Ann Surg 172:787-792, 1970. 8. Sparks CH: Silicone mandril method of femoropopliteal artery bypass.
Am J Surg 124:244-249, 1972. 9. Morgan AP, Lazarus JM: Vascular access for dialysis: Techniques and newer methods. Am J Surg, to be published. 10. Sauvage LR, Berger K, Wood SJ, et al: An external velour surface for porous arterial prosthesis. Surgery 70:940-953, 1971. 11. Brais MP, Braunwald NS: Tissue acceptance of materials implanted within the circulatory system. Arch Surg 109:351-358, 1974. 12. Ghidoni JJ, Liotta D, Hall CW, et al: Healing of pseudointimas in velour-lined, impermeable arterial prosthesis. Am J Pathol 53:375-390, 1968. 13. Wesolowski SA, Fries CC, Karlson KE, et al: Porosity: Primary determinant of ultimate fate of synthetic vascular grafts. Surgery 50:91-96, 1961. 14. Hern\l=a'\ndez-Ja\l=u'\reguiP, Esperanza-Garci\l=a'\C, Gonz\l=a'\lez-AnguloA: Morphology of the connective tissue grown in response to implanted silicone rubber. A light and microscopic study. Surgery 75:631-637, 1974. 15. Florey HW, Greer SJ, Kiser J, et al: The development of the pseudointima lining fabric grafts of the aorta. Br J Exp Pathol 43:655-660, 1962. 16. Gillman T, Wright LJ: Autoradiographic evidence suggesting in vivo transformation of some blood mononuclears in repair and fibrosis. Nature 209:1086-1090, 1966.
results with
Downloaded From: http://archsurg.jamanetwork.com/ by a Osaka University User on 05/14/2015
