Genetic tracing of arterial graft flow surface endothelialization in allogeneic marrow transplanted dogs.

In order to trace genetically the source of fallout endothelialization on arterial grafts, six beagle dogs with successful autologous bone marrow transplantation received composite tandem aortic grafts with an isolated, totally impervious Dacron graft and a porous Dacron graft for 12 weeks. For impervious segments, five of 12 fresh tissue samples were Factor VIII/von Willebrand factor + (FVIII/vWF) and seven had faint or negative signals; three of the FVIII/vWF + samples had alpha-actin + smooth muscle cells. Polymerase chain reaction (PCR) study showed eight had a pure donor DNA genotype and four had donor/host mixed, with the donor predominant. Of 12 AgNO3-stained samples, 11 showed pure donor type and one had donor/host mixed, with the donor predominant. For porous segments, all 12 fresh samples had positive flow surface FVIII/vWF and alpha-actin cells. PCR showed all these samples and all 12 AgNO3-stained samples had donor/host mixed type, but the host pattern was predominant. Porous graft healing appears to involve both cellular fallout and tissue ingrowth, and bone-marrow-derived cells may be a source for fallout.

The effect of flow shear stress on endothelialization of impervious Dacron grafts from circulating cells in the arterial and venous systems of the same dog.

The purpose of this report was to study effects of shear force and hemodynamic conditions that influence fallout healing in the arterial and venous systems of the same dog. Knitted Dacron grafts made impervious by a 1.5 mm thick coat of silicone rubber bonded to the external surface were implanted for 4 weeks during the same surgery in the descending thoracic aorta (DTA), abdominal aorta (AA) and inferior vena cava (IVC) of each of five dogs. Flow rates were measured during surgery and shear stresses calculated with the Hagen-Poiseuille formula. Full-wall thickness longitudinal tissue sections were embedded in resin and stained with hematoxylin and eosin for light microscopy, and in paraffin for immunocytochemistry studies with Factor VIII/von Willebrand factor, smooth muscle alpha-actin, collagen IV, laminin, and proliferating cell nuclear antigen. Scanning electron microscopy and transmission electron microscopy studies were also performed. AgNO3 was used to determine percentage of endothelial-like cell coverage on the flow surface. All grafts were patent, without hematoma or seroma. Endothelial-like cell coverage was highest in the IVC grafts and lowest in the DTA. Shear stress and flow velocity were significantly lower in IVC grafts than DTA and AA. Proliferating cell nuclear antigen indicated extensive cellular proliferation in the intima and in the interstices of the inner portion of the graft wall. The degree of fallout healing in knitted Dacron grafts made impervious by an external coat of silicone rubber varies inversely with the sheer force of blood flow in these grafts.

Early flow surface endothelialization before microvessel ingrowth in accelerated graft healing, with BrdU identification of cellular proliferation.

The purpose of this report was to determine if flow surface endothelialization could precede microvessel ingrowth from the perigraft area in porous Dacron grafts, by using an accelerated graft healing model with short implant periods. Dacron grafts were implanted in the abdominal aorta of 22 dogs and wrapped in autogenous inferior vena cava (IVC), which provided excellent conditions for extramural angiogenesis, microvessel development, and ingrowth toward the graft. Retrieval times were 7 days (n = 4), 8 days (n = 5), 9 days (n = 4), 10 days (n = 3), 11 days (n = 4) and 12 days (n = 3) postoperatively. Graft surfaces were evaluated for thrombus coverage, cell coverage, and the number of micro-ostia. Components and cellular types in the graft wall and on the surface were studied and characterized with H&E, histochemical, and immunocytochemical staining. BrdU labeling was also used, to identify the areas where cells were actively proliferating. All grafts were patent. Although the degree of IVC/graft attachment varied, isolated islands of endothelial-like cells were found at the midgraft areas at each time period, and immunocytochemically confirmed as endothelial cells. There were two healing patterns: (1) surface endothelialization before microvessel/tissue ingrowth from the perigraft areas, and (2) surface endothelialization with full wall microvessel and tissue presence. Surface endothelialization was observed before perigraft tissue ingrowth, indicating that fallout healing is an independent source of endothelialization for porous grafts.

Apparent blood stream origin of endothelial and smooth muscle cells in the neointima of long, impervious carotid-femoral grafts in the dog.

The purpose of this study was to determine whether endothelial and smooth muscle cells originating from the blood stream contribute to the endothelialization of impervious, small-caliber, long Dacron grafts used as extraanatomical bypasses in dogs. We implanted silicone-rubber-coated, permanently impervious grafts 64 to 77 cm long and 6 mm in diameter, made of externally supported knitted Dacron as unilateral carotid-femoral bypasses with distal femoral arteriovenous fistulae in 10 dogs for 3 months; sides were alternated between cases. Subjects received 162 mg/day of aspirin, and its effectiveness on platelet aggregation (PA) was evaluated and expressed as a PA score. Graft healing was studied by stereomicroscopy with silver nitrate staining, by light microscopy with hematoxylin-eosin and immunocytochemical staining for endothelial and smooth muscle cells, and by scanning and transmission electron microscopy. Five grafts were patent for 3 months and could be included in the healing study; the five occluded grafts thrombosed within 14 days. Although there was no transinterstitial tissue ingrowth from perigraft tissues into the impervious Dacron grafts, scattered islands of endothelial cells were conclusively demonstrated on graft flow surfaces 3 months after implantation. Average endothelial-like cell coverage of the flow surfaces was 15.6% +/- 3.8%, and alpha-actin-positive smooth muscle cells and microvessels were found beneath some of the endothelial islands. These findings suggest that blood stream-derived endothelial and smooth muscle cells play a role in the healing of the inner wall of Dacron grafts in the dog.

Differential effect of the retropleural and retroperitoneal environments on healing of the inner wall of porous fabric prostheses in the thoracic and abdominal aorta of the same dog.

Healing of the inner wall of the same preclotted knitted Dacron arterial prostheses was compared in the descending thoracic aorta (DTA) and the abdominal aorta (AA) of the same dog. Each of 16 dogs received this dual implantation with study periods of 4 weeks for five dogs, 8 weeks for five dogs, and 16 weeks for six dogs. Healing was studied with light microscopy, scanning electron microscopy, transmission electron microscopy, and immunocytochemistry identification. The outer capsule was firmly adherent to the Dacron framework of all grafts implanted in the DTA; advanced healing of the inner wall of all thoracic grafts was present by 4 weeks, nearly complete healing by 8 weeks, and complete healing by 16 weeks. In contrast, the outer capsule was either not attached or only loosely adherent to the Dacron framework in eight AA grafts (50%), and in these implants no healing of the inner wall occurred beyond the limited perianastomotic pannus zone. In the other eight implants in which the outer capsule was firmly adherent to the Dacron framework, healing was roughly comparable to that in the grafts implanted in the DTA. This study demonstrated that (1) DTA grafts heal faster and more completely than AA grafts, (2) healing and endothelialization are related to the tightness of the outer capsule, (3) there is a high incidence of loose tissue attachment in the AA, and (4) healing of aortic grafts is site specific.

Definitive proof of endothelialization of a Dacron arterial prosthesis in a human being.

A 10 mm woven Dacron axillofemoral bypass graft was removed from a 65-year-old patient during redo surgery after an implant period of 26 months, because of a large seroma that surrounded the entire length of the graft. Tissue blocks were taken from representative areas along the entire length of the graft surface and evaluated by light microscopy with hematoxylin and eosin and Masson trichrome staining, scanning electron microscopy, transmission electron microscopy, and immunocytochemical staining. Paraffin-embedded sections were stained with smooth muscle cell alpha-actin, which demonstrated smooth muscle cells in the pseudointima, and Ham 56 stain to identify macrophages. Endothelial factor VIII/von Willebrand factor and Ulex europaeus agglutinin identified human endothelial cells on the flow surface, in areas far removed from the anastomoses to the native vessels. This is the first definitive proof in a human of endothelialization of a synthetic arterial graft beyond the pannus ingrowth zone.

Proof of fallout endothelialization of impervious Dacron grafts in the aorta and inferior vena cava of the dog.

PURPOSE: The purpose of this study was to prove directly whether cells from the blood stream contribute to endothelialization of isolated, impervious Dacron vascular grafts in the dog. METHODS: We designed an 18 cm, three-component graft with two parallel central Dacron limbs; one was made impervious with silicone rubber, and the other was preclotted. This model was implanted in the canine descending thoracic aorta with 30 microns polytetrafluoroethylene grafts anastomosed at each end. An 8 cm, three-component graft completely coated with silicone rubber was implanted in the canine abdominal aorta and inferior vena cava. Implant periods ranged from 4 to 12 weeks. Flow surfaces were studied by use of stereomicroscopy after being stained with silver nitrate, and by use of scanning and transmission electron microscopy, the inner wall and flow surface were studied by light microscopy after hematoxylin-eosin and immunocytochemical staining (the latter for endothelial and smooth muscle cells), and the full wall was studied by light microscopy after hematoxylin-eosin staining. RESULTS: Effective prevention of pannus and transmural ingrowth into the impervious central test grafts was achieved, and scattered islands of endothelial cells were conclusively demonstrated on flow surfaces in each of the three implant sites 4 weeks after implantation. In the descending thoracic aorta, where these grafts were also implanted for 8 and 12 weeks, alpha-actin-positive cells and microvessels were found beneath some of the endothelial islands. CONCLUSION: Fallout endothelialization of Dacron vascular grafts occurs in both the arterial and venous systems of the dog.

Mitosis and cytokinesis in subconfluent endothelial cells exposed to increasing levels of shear stress.

An in vitro flow apparatus in combination with cultured endothelium was used to determine the effects of fluid-generated shear stress on cells undergoing mitosis and cytokinesis. Cell responses were recorded by time-lapse video microscopy under phase contrast or Hoffman modulation contrast optics. Completion of cell division in mitotic cells was dependent upon both the initial presence of intercellular attachments and the magnitude of fluid wall shear stress. In nonisolated populations, 95.3%, 69.5%, and 57.1% of the cells completed cell division as opposed to 66.6%, 20.4%, and 11.9% in the isolated cell groups at 2.8, 14.1, and 33 dynes/cm2, respectively. Prestressing cells for 14 h prior to monitoring failed to increase retention of isolated mitotic cells. The presence of neighboring cells facilitated replication by providing an anchoring attachment or a luminal surface for completion of division. Cell detachment most commonly occurred at the onset of cytokinesis when substrate contact areas were minimal and focal contacts were absent. A comparison between no flow controls and shear stress specimens indicated no significant differences in transit times for mitosis and cytokinesis. Thus, subconfluent endothelial cells may be more susceptible to detachment during cell division due to increases in shear stress, the absence of intercellular attachments, and reduced cell-substrate contacts.

Effect of shear stress upon localization of the Golgi apparatus and microtubule organizing center in isolated cultured endothelial cells.

Despite substantial evidence to suggest that directed cell migration is dependent upon positioning of the Golgi apparatus (GA) and the microtubule organizing center (MTOC), some controversy exists about whether such a relationship is relevant to endothelial cells under flow. The present study was undertaken to provide an indepth investigation of the relationship between shear stress, GA/MTOC localization, cell migration and nuclear position. Bovine carotid endothelial cells were exposed to 22 or 88 dynes/cm2 for 0.5, 2, 8 or 24 h, and localization of their GA/MTOCs was determined relative to the direction of flow. In no-flow control specimens, (0, 0.5, 2, 8 and 24 h) there was no change in the equally distributed GA/MTOCs. In contrast, during the first 8 h at 88 dynes/cm2 and by 2 h at 22 dynes/cm2 there was a significant increase in the number of cells with GA/MTOCs localized upstream to flow direction. The effect was temporary, however, and by 24 h there was no significant difference between the no-flow, 22 and 88 dynes/cm2 specimens. Analysis of GA/MTOC localization with respect to the direction of cell migration determined that 72.5% of no-flow cells possessed GA/MTOCs localized to the sides of nuclei nearest the direction of migration. In contrast, 64% of the specimens shear stressed over the same time period had GA/MTOCs localized to the sides of nuclei opposite the direction of migration. These results suggest that positioning of the GA/MTOC in endothelial cells is not dependent completely upon the direction of migration.(ABSTRACT TRUNCATED AT 250 WORDS)

The direct effect of graft compliance mismatch per se on development of host arterial intimal hyperplasia at the anastomotic interface.

To study the direct and sole effect of compliance mismatch on anastomotic intimal hyperplasia of the host arterial wall and to minimize possible confounding factors, dogs with a low thrombotic potential were selected as experimental subjects. Externally supported 6 cm x 5 mm Dacron grafts with a compliance value of approximately 1/300 of the host artery were implanted into the carotid arteries with end-to-end anastomoses on one side and end-to-side anastomoses on the other. The control graft was an autogenous carotid artery segment 4 cm in length transplanted into the femoral artery. Eight cases (24 grafts) were studied for 1 year and three (nine grafts) for 6 months. All were patent throughout the study period except for two noncompliant grafts with end-to-end anastomoses; thrombosis was the documented cause of occlusion. For the patent grafts, follow-up arteriograms showed no progressive narrowing of noncompliant anastomoses. Whether compliant or noncompliant, light microscopy studies showed slight intimal thickening within 1 to 2 mm of the anastomotic line, possibly the result of the normal healing response to stitch and surgical trauma. Quantitatively, 22 measurements representing longitudinal and circumferential thickness of the neointima were taken at each of the 40 patent noncompliant and 22 patent compliant control anastomoses. There was no statistically significant difference in anastomotic neointimal thickness in compliant and noncompliant grafts or for the different implantation periods. These data suggest that graft/host artery compliance mismatch does not cause arterial intimal hyperplasia at the anastomotic interface.

Dextran increases survival of subconfluent endothelial cells exposed to shear stress.

In vitro devices in combination with cultured cells have been used to study the relationship between shear stress and endothelial injury. Almost exclusively, these investigations have used confluent monolayers and conventional culture media as perfusates and reported little cell loss over a wide range of shear stress conditions. In this investigation when subconfluent endothelial cells were exposed to 22 and 88 dyn/cm2 for 2, 8, and 24 h in a perfusate of medium and 5% serum, a progressive cell loss was observed. Lower cell densities were a product of decreased cell proliferation as measured by bromodeoxyuridine (BrdU) incorporation and loss of the initial cell population. Video recordings indicated that cells characteristically detached by proximal cell peeling from the substrate and an aneurysmal rupture of the cell membrane. Cell retention was increased by including 250 and 475 microM neutral dextran (70 kDa) in perfusates. Experimental evidence suggests dextran does not directly stimulate proliferation or correct an osmotic imbalance. This investigation has substantiated that fluid-generated shear stress can cause endothelial denudation and that conditions (subconfluency, time, and perfusate supplementation) under which shear stress is applied are as important for cell survival as shear stress magnitude.

Experimental evaluation of ten clinically used arterial prostheses.

The arterial grafts currently in use are classified into five basic categories; 10 different commercially available prostheses were chosen to represent these categories. The Vascutek and Bionit are made from knitted Dacron and have medium porosity, requiring preclotting by the surgeon. The DeBakey Soft Woven and Plasma TFE grafts are made from woven Dacron and have low porosity, not requiring preclotting under most circumstances. Also studied were woven and knitted grafts with leakage resistance referred to as "impervious" Dacron grafts: the Vascutek Gelseal, the Bard Albumin Coated DeBakey Vasculour II, the Microvel with Hemashield, and the albumin saturated, autoclaved DeBakey Soft Woven graft. Gore-Tex and Impra are expanded polytetrafluoroethylene grafts which do not require preclotting. For each type, five grafts 6 cm long and 8 mm in diameter were implanted in the descending thoracic aorta of healthy adult dogs for 16 weeks. The physical characteristics, biocompatibility, and healing patterns varied according to the structure and treatment of the grafts. Pretreatment with biomaterials during manufacture is quite effective in preventing transinterstices blood loss during implantation, but results in altered physical qualities, increased thrombogenicity and delayed healing in comparison to the effects of preclotting with autogenous blood at the time of implantation.

Re-endothelialization of isolated segments of the canine carotid artery with reference to the possible role of the adventitial vasa vasorum.

This study was designed to determine if the adventitial vasa vasorum contribute to re-endothelialization of the canine carotid artery after removal of the endothelial flow surface. Casting studies demonstrated that vasa vasorum are present only in the adventitia of the canine carotid artery. Ninety autograft segments of the carotid artery from which the endothelium had been removed were implanted in both carotid arteries of 45 dogs. Glutaraldehyde-processed canine carotid allografts were positioned at each end to prevent pannus ingrowth and a Gore-Tex wrap to prevent periarterial tissue growth was placed into the outer wall. Three methods were used, with observations at 4 and 8 weeks. In method 1 the test segment was treated with superficial endarterectomy. In method 2 a balloon catheter was used to remove the endothelium. In method 3 balloon catheter denudation of the flow surface was also employed and, in addition, the adventitia was removed surgically as completely as possible, although a few vasa vasorum were shown to remain in some grafts. In method 1 all the patent endarterectomized arteries were partially re-endothelialized at both the 4- and 8-week intervals. In method 2, 72% of the balloon-denuded patent arteries with intact adventitial vasa vasorum were partially re-endothelialized at 4 weeks and 84% at 8 weeks. However, in method 3 the flow surfaces had no endothelium at 4 weeks and 83% still had none at 8 weeks. These findings suggest that, in the absence of pannus ingrowth, re-endothelialization of the canine carotid artery depends on not only the presence but also the number of adventitial vasa vasorum.

Implant site: a determinant of completeness of arterial prosthesis healing in the dog and possibly in humans.

This paper compares the healing of supported knitted Dacron prostheses implanted in the descending thoracic aorta and in the subcutaneous carotid-femoral positions in each of 10 dogs. The descending thoracic aorta prostheses were 6 cm long and the carotid-femoral prostheses averaged 76 cm in length. The healing of four porous clinical axillofemoral grafts is compared to that of the experimental carotid-femoral grafts. There was little similarity between the healing observed in the descending thoracic aortic and subcutaneous positions; the descending thoracic aorta grafts healed rapidly and completely by eight weeks, with thrombus-free flow surfaces covered with confluent endothelial cells. However, the inner wall of the carotid-femoral grafts remained largely unhealed at one year, with endothelialization largely confined to a limited pannus ingrowth across the anastomoses. This indicates that the implant site in the dog is a primary determinant of the rapidity and completeness of healing. The clinical axillofemoral grafts explanted after 20 months to six and one-half years exhibited healing characteristics that were similar to those seen in the carotid-femoral position in the dog. These findings suggest that valid comparisons of interspecies healing require data obtained from similar implant locations. They present a challenge to the previously held concepts of major interspecies differences in speed and extent of tissue ingrowth based on data obtained from noncomparable implant sites.

The evolution of a clinically relevant experimental model for the realistic evaluation of long-length (80-cm) arterial prostheses.

A new carotid-femoral experimental model has been developed which has significant clinical relevance to human limb salvage bypass grafting in the areas of graft length, hemodynamics and tissue environment. More significantly, the graft healing pattern observed with this model is similar to what is seen in humans. Therefore, this experimental model provides a challenging and critical site for dynamic healing studies and comparisons of different types of vascular prostheses.

Endothelial adherence under shear stress is dependent upon microfilament reorganization.

In response to externally applied shear stress, cultured endothelial monolayers develop prominent, axially-aligned, microfilamentous bundles, termed "stress fibers" (Dewey: Journal of Biomechanical Engineering 106:31-35, 1984; Franke et al.: Nature 81:570-580, 1984; Franke et al.: Klin. Wochenschr 64:989-992, 1986; Wechezak et al.: Laboratory Investigation 53:639-647, 1985). It is unclear, however, whether similar stress fibers develop in noncontiguous endothelial cells and whether these structures are necessary for adherence of individual cells under shear stress. It also is unknown what alterations occur in microtubules, intermediate filaments, and focal contacts as a consequence of shear stress. In this study, endothelial cells, free of intercellular contact, were exposed to 93 dynes/cm2 for 2 hr. With the aid of specific labeling probes and interference reflection microscopy, the distributional patterns of microfilaments, microtubules, intermediate filaments, and focal contacts were examined. Following shear stress, microfilament bundles and their associated focal contacts were concentrated in the proximal (relative to flow direction) cell regions. In contrast, microtubules were distributed uniformly within cell contours. Intermediate filaments displayed only an occasional tendency for accumulation at proximal edges. When cells were shear-tested in the presence of cytochalasin B to inhibit microfilament assembly, considerable cell loss occurred. Following inhibition of tubulin polymerization, no increase was observed in the percentage of cells lost due to shear over nontreated controls. Nocodazole-treated cells, however, were characterized by prominent stress fibers throughout the cell. These results indicate that stress fiber and focal contact reorganization represent major responses in isolated endothelial cells exposed to shear stress and that these cytoskeletal structures are necessary for adherence.

Differences in reendothelialization after balloon catheter removal of endothelial cells, superficial endarterectomy, and deep endarterectomy.

The process of reendothelialization was studied in a deendothelialized 3 cm segment of the canine descending thoracic aorta from which peripheral endothelial cell ingrowth had been prevented by impervious prosthetic graft sleeves. Three preparations were studied: (1) removal of only the endothelial cells, accomplished by flow surface drying and balloon catheter denudation, (2) removal of the superficial portion of the nonvasal media, accomplished by endarterectomy to a depth of 200 to 300 micron, and (3) removal of the entire inner (nonvasal) media, accomplished by endarterectomy to a depth of 500 to 600 micron to reach the outer (vasal) media. A total of 44 specimens were studied after implantation periods of 7, 14, 28, and 56 days. In all cases endothelial cell ingrowth from the host aorta into the test specimen was prevented by the impervious graft sleeves. In the deep endarterectomy group there were scattered areas of reendothelialization at 1 week, extensive reendothelialization at 2 weeks, almost complete reendothelialization at 4 weeks, and confluence by 8 weeks. However, in the superficial endarterectomy group scanning electron microscopy showed scattered areas of factor VIII/von Willebrand factor (FVIII/vWF)-negative, endothelial-like cells at 4 weeks, whereas at 8 weeks most of the surface was covered by endothelial cells identified by FVIII/vWF. In those specimens subjected to balloon catheter removal of endothelial cells only, reendothelialization was not seen, even at 8 weeks.

A rapid in vitro test of the in vivo healing potential of vascular prostheses.

An in vitro method for comparing the penetration of bovine fibroblasts seeded on the external surface of porous vascular prostheses was devised. The effects of water porosity reduction and differing manufacturing constructions (warp-knit Dacron, woven Dacron and polytetrafluoroethylene) on the ability of the bovine fibroblasts to penetrate transmurally was investigated. Of the warp-knit external-veloured Dacron prostheses, the highest porosity 140-denier prototype had the highest luminal surface cell count and the lowest porosty 280-denier prototype the lowest luminal surface cell count. The intermediate prototypes had values between these two extremes. The woven Dacron prostheses, which were of even lower porosity but with a much thinner wall, had cell counts midway between the 140-denier and the 280-denier prototypes. The microporous polytetrafluoroethylene prostheses did not allow fibroblast penetration despite adherence of cells to the outer surface. These findings agree with in vivo healing studies of the same materials in the descending thoracic aorta of the dog, demonstrating that this rapid in vitro assay method can help predict the healing potential of a vascular prosthesis.

An apparatus to study the response of cultured endothelium to shear stress.

An apparatus which has been developed to study the response of cultured endothelial cells to a wide range of shear stress levels is described. Controlled laminar flow through a rectangular tube was used to generate fluid shear stress over a cell-lined coverslip comprising part of one wall of the tube. A finite element method was used to calculate shear stresses corresponding to cell position on the coverslip. Validity of the finite element analysis was demonstrated first by its ability to generate correctly velocity profiles and wall shear stresses for laminar flow in the entrance region between infinitely wide parallel plates (two-dimensional flow). The computer analysis also correctly predicted values for pressure difference between two points in the test region of the apparatus for the range of flow rates used in these experiments. These predictions thus supported the use of such an analysis for three-dimensional flow. This apparatus has been used in a series of experiments to confirm its utility for testing applications. In these studies, endothelial cells were exposed to shear stresses of 60 and 128 dynes/cm2. After 12 hr at 60 dynes/cm2, cells became aligned with their longitudinal axes parallel to the direction of flow. In contrast, cells exposed to 128 dynes/cm2 required 36 hr to achieve a similar reorientation. Interestingly, after 6 hr at 128 dynes/cm2, specimens passed through an intermediate phase in which cells were aligned perpendicular to flow direction. Because of its ease and use and the provided documentation of wall shear stress, this flow chamber should prove to be a valuable tool in endothelial research related to atherosclerosis.

The influence of denier and porosity on performance of a warp-knit Dacron arterial prosthesis.

The study objective was to determine the optimal water porosity for a warp-knit, externally veloured Dacron arterial prosthesis. Preclotting efficiency, surgical handling characteristics, thrombogenicity, and 56-day healing in the descending thoracic aorta of mongrel dogs were measured for seven different arterial prostheses of the same basic design constructed to test the influence of decreased porosity produced by increased yarn denier at a constant needle density. The six prototypes tested were 160, 180, 200, 220, 240, and 280 denier. The 140 denier prosthesis (Bionit, C. R. Bard, Inc., Billerica, MA) served as the control. The water porosity of these prostheses varied from 2036 ml/cm2/min at 120 mmHg for the control to 286 for the 280 prototype. The preclotting efficiency increased with decreased water porosity and with increased outer-wall filamentousness. Desirability of surgical handling characteristics decreased with lowered water porosity, while the degree of outer-wall filamentousness had little influence. All the prostheses showed excellent through-wall healing at 56 days, including endothelialization. The 160 prototype represented an optimal blend of high preclotting efficiency, desirable surgical handling characteristics, and excellent 56-day healing. These studies are pertinent to the future development of porous, warp-knit Dacron arterial prostheses.