Reduced Susceptibility of Polytetrafluoroethylene Vascular Prostheses to Colonization by Staphylococcus aureus Following In Situ Endothelialization

Abstract: The impact of endothelialization of polytetrafluoroethylene (PTFE) grafts on susceptibility to experimental colonization by Staphylococcus aureus was studied in a rat model. One hundred and eight grafts (2 mm inner diameter, 5 mm length) were implanted into the infrarenal aorta (54 rats) or the infrarenal caval vein (54 rats). The progress of endothelialization following graft implantation was evaluated by SEM at 1, 3, 7, and 14 days on 6 grafts from each group. We found that the endothelialization was more rapid in caval vein than in aorta: The caval vein grafts were completely endothelialized 2 weeks following implantation compared with endothelialization of approximately 0.5 mm of aorta grafts, measured from each anastomosis. During this time, the resistance to standardized intravenous challenge with 10⁸ colony-forming units of S. aureus increased gradually in both groups, and all caval vein grafts tested at 2 weeks were sterile. However, all aorta grafts challenged at 2 weeks were colonized, although to a somewhat lower degree than at earlier challenge. Resistance of PTFE grafts to colonization with S. aureus thus correlated to the degree of endothelialization.     

Prosthesis replacement of the inferior vena cava

The development of reconstructive venous surgery has been hampered by the lack of suitable graft materials. Fourteen carbon fibre grafts (phi: 8 mm), 30 glutaraldehyde treated ovine collagen grafts and 29 polytetrafluoroethylene (PTFE) grafts were used to replace a 35 mm segment of infrarenal inferior vena cava in pigs. Prostheses were removed 1 hour and 7, 14, 28, 56 and 112 days after implantation. All specimens were examined by light and scanning electron microscopy. The 112 day patency rate was 67% for ovine collagen grafts, while all carbon and PTFE grafts thrombosed. The difference was statistically significant (p less than 0.01). During the first hour after implantation, a thick (800-900 microns) thrombotic layer deposited on the inner surface of carbon grafts. This layer possibly caused the subsequent complete occlusion of the tubular segments. A thin neointima (less than 200 microns) developed on the flow surface of ovine collagen prostheses. This favoured complete endothelialization of the graft inner surface as soon as four weeks after surgery. In conclusion, glutaraldehyde treated ovine collagen would represent the first sound material to be used as venous substitute.     

Experimental microvenous reconstructions with Gore-Tex polytetrafluoroethylene prosthesis implanted by means of the sleeve anastomotic technique

Polytetrafluoroethylene (PTFE) prostheses (Gore-Tex; ID, 1 mm; length, 5-7 mm; wall thickness, 0.2 mm; fibril length, 30 microns, n = 28) were implanted into the rat femoral vein by means of the sleeve anastomotic technique to enhance the patency rate. In the control group, PTFE prostheses (n = 8) were implanted by means of the end-to-end technique. In the experimental group patency and healing of the PTFE prostheses were evaluated at 1 day (n = 4), 1 week (n = 6), 3 weeks (n = 6), 6 weeks (n = 6), and 12 weeks (n = 6) after implantation by means of macroscopic inspection and routine light and scanning electron microscopy. All prostheses, except one at 1 week after implantation, were patent at the time of removal. All of the microvenous prostheses were completely covered by an endothelial layer at 3, 6, and 12 weeks after implantation. Occasionally some smooth muscle-like cells could be found underneath this endothelial layer, but stenosis was never observed at the anastomotic sites. Only scarce tissue ingrowth was observed in the wall of the PTFE prostheses. In the control group, all prostheses, except one prosthesis after 3 weeks, were found to be occluded. An occlusive mural thrombus was found firmly attached at the anastomoses at 1 day, and an organized thrombus at 3 weeks after implantation. The patent prosthesis demonstrated complete endothelial healing. These results demonstrate the importance of the sleeve anastomotic technique and the potential of PTFE prostheses as a microvenous conduit when implanted by means of the sleeve anastomotic technique in experimental reconstructive microvascular procedures.     

The influence of inner wall filamentousness on the performance of small- and large-caliber arterial grafts

Dacron prostheses with differing degrees of inner wall filamentousness were implanted in the canine descending thoracic aorta for 56 days (n = 14) and the carotid arteries for 21 days (n = 40). In the thoracic aorta the healed inner capsule of 8 mm internal velour grafts was approximately three times as thick as the inner capsule of external velour knitted prostheses. However, the early patency and rapidity of healing of these two types of prostheses were essentially the same. In the carotid arteries the patency of the 4 mm external velour prostheses was greater than that of the internal velour prostheses and the mean thrombus-free surface area of the external velour grafts was significantly greater than that of the internal velour grafts. These observations suggest that an internal velour surface is not deleterious to early patency in a large-caliber, high-flow artery but that it may be in a small-caliber, low-flow artery.     

Composite porosity of expanded polytetrafluoroethylene vascular prosthesis

Polytetrafluoroethylene (PTFE) prostheses were modified to produce two types of composite porosity PTFE grafts: type I — inner 60 μm/outer 20 μm and type II — inner 20 μm/outer 60 μm. These composite porosity PTFE designs were investigated for bleed through, graft healing and patency and compared with high porosity PTFE (60 μm) and standard PTFE (20 μm) grafts. The grafts were implanted into the carotid and femoral arteries of dogs and retrieved after 4, 12 and 18 weeks. Both composite porosity grafts showed significantly less bleed through than standard and high porosity grafts after reperfusion. In composite grafts, the 60-μm layer allowed fibrovascular tissue and histiocyte ingrowth from perigraft tissue, but the 20-μm layer did not. Neointima formation occurred earlier and endothelialization was more extensive in high porosity grafts, but seroma formation occurred in 25% of cases. In type I composite porosity grafts, smooth muscle cells of neointima migrated into the pores of the graft, providing a firmly anchored intima. Type II composite porosity grafts allowed better ingrowth of fibrovascular tissue at the outer layer from the perigraft tissue; however, endothelialization was not completed. Composite porosity grafts should be considered for evaluation in specific clinical situations.     

Platelet antagonists eliminate thromboembolic complications of small-diameter arterial prostheses

An animal model suitable for study of the origin and method of prevention of thromboembolic complications of arterial prostheses has been developed in the rabbit. In phase I of the experiments 42 New Zealand white rabbits underwent insertion of polytetrafluoroethylene (PTFE) aortic grafts, 10 mm in length and of 2 mm internal diameter (I.D.) (n = 17) and 3 mm I.D. (n = 25). The patency rate at 3 months was 24% and 82%, respectively. Unexpected ischemic hind limb ulcers occurred in nine (38%) of the long-term survivors. Arteriograms in these animals showed a typical embolic occlusion of a distal artery, suggesting that the ulcers were due to embolization of loose mural thrombus, which was present in 50% of the grafts when removed. In phase II experiments 54 rabbits were randomly allocated to receive aspirin (ASA) 10 mg/kg/day and dipyridamole (DPM) 10 mg/kg/day (n = 25) or placebo (n = 29). Both regimens began 3 days before insertion of PTFE aortic grafts (10 mm long and 3 mm I.D.). Serum thromboxane B2 concentrations in the control group averaged 300.4 +/- 147.4 ng/ml and 43.2 +/- 58.6 ng/ml in the ASA/DPM group (p less than 0.0005). With the use of autologous indium 111 oxine-labeled platelets, a graft platelet accumulation index (GPAI) was calculated as the graft: reference ratio of emissions. ASA/DPM significantly reduced the mean GPAI calculated from grafts and reference aorta removed 48 hours after graft insertion from 69.3 +/- 4.0 on placebo (n = 4) to 34.3 +/- 2.9 (n = 4) (p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)     

Neointima formation in expanded polytetrafluoroethylene vascular grafts with different fibril lengths following implantation in the rat aorta

Expanded polytetrafluoroethylene (PTFE) prostheses with fibril lengths of 30 and 60 microns were implanted in the rat infrarenal aorta. Sequential scanning electron and light microscopic studies of the prostheses after implantation demonstrated a different pattern of endothelialization. Prostheses with a fibril length of 60 microns had a continuous multilayered neointima at week 25 postimplantation, whereas prostheses with a 30-microns fibril length had a discontinuous and single layer of endothelium after the same interval. It was concluded, therefore, that a prerequisite for the development of a lining in a vascular prosthesis is for the inner surface of the prosthesis to have adequate pores for effective anchoring of the invading endothelioid cells. Expanded PTFE prostheses with an internodular distance of 60 microns provided sufficient anchoring possibilities for invading endothelioid cells to form a continuous neointima.     

Preservation of compliance in a small diameter microporous, silicone rubber vascular prosthesis

Compliance is believed to be a significant factor in maintaining the patency of small diameter vascular grafts. This study evaluated the compliance changes with time of microporous Replamineform silicone rubber prostheses. The compliance of 15 canine femoral artery interpositions (4 mm internal diameter X 6 cm length) was measured by in-vivo electromagnetic rheoangiometry immediately following implantation and at intervals to eight months. At implantation, silicone rubber grafts were overcompliant (15.0 +/- 1.1% radial change/mmHg X 10(-2); mean +/- S.E.) compared to the proximal artery (7.7 +/- .6%). The compliance of the prostheses decreased within two weeks (6.3 +/- .9%) and remained isocompliant to the proximal artery for eight months. The compliance determinations for the silicone rubber grafts were compared with those from PTFE prostheses and vein grafts acquired by the same method in a previous study. The analysis demonstrates the preservation of isocompliance of the silicone rubber prostheses compared to the native arteries. In contrast, the minimally compliant vein grafts and PTFE prostheses continued to decrease in compliance following implantation. This microporous silicone rubber graft may improve the success of small internal diameter arterial reconstructions by eliminating failures caused by compliance mismatch between the artery and the prosthesis.     

Effect of hemodynamic conditions and prosthesis structure on endothelialization of clinical and experimental small lumen vascular prostheses]

Seeding of small-diameter vascular prostheses (ID less than or equal to 6 mm) with autologous microvascular cells (AMVC) results in a complete endothelial cell layer on the luminal surface. The purpose of this study was to examine the influence of the blood flow velocity (due to 4 or 6 mm ID) and the structure of inner graft surface (crimped, uncrimped) on the endothelialization. AMVC were harvested from omental adipose tissue (mean: 0.56 X 10(6) cells/g tissue) from 10 mongrel dogs (mean: 27.9 kg). During preclotting, the 4 mm uncrimped and the 6 mm crimped double velour Dacron prostheses (Meadox Medicals, Inc.) were seeded with 1.0 X 10(6) cells/cm2 graft surface. Grafts were implanted into the carotid arteries (N = 5 in each group). The animals received antiplatelet therapy. After five weeks, all seeded prostheses were patent. The thrombus free surface (TFS) of seeded prostheses was 99.9% (4 mm) and 90.5% (6 mm). Scanning electron microscopy revealed an athrombogenic layer of endothelial cells on a smooth surface. -It is concluded that in canine experiments endothelialization of 4 and 6 mm grafts after seeding with AMVC is not affected by blood flow velocity or graft structure.     

Rapid endothelialization of vascular prostheses by seeding autologous venous tissue fragments.

A method was developed to obtain rapid endothelialization of a fabric vascular prosthesis by seeding autologous venous tissue fragments into its wall. In an animal study, complete endothelialization was observed in the entire inner surface of the prosthesis within 2 weeks after implantation. A piece of peripheral vein was minced with scissors and then stirred into saline to create a tissue suspension. This suspension was enmeshed into the wall of a highly porous fabric vascular prosthesis by repeated pressurized injections with a syringe. The prostheses (7 mm inside diameter and 5.7 cm in length), seeded with tissue fragments, were implanted into the descending thoracic aorta of 25 dogs, and they were removed from 1 hour to 2 months after implantation. Twenty-five prostheses, preclotted with fresh blood, were used as control prostheses. In the seeded graft, a thin fibrin layer covered the inner surface just after implantation, but countless numbers of endothelial cells migrated from the fragments and came up to the luminal surface like multiple "mushrooms" under the fibrin layer. Smooth muscle cells made multiple layers underneath the endothelial cell layer. The healing proceeded equally at every part. By this active migration and proliferation, the inner surface was completely healed within 2 weeks.     

Glutaraldehyde-tanned mandril-grown grafts as venous substitutes

The present study was performed to evaluate the potential of glutaraldehyde-tanned mandril-grown grafts as caval substitutes. Short-term experiments consisted of 30 tubular grafts (35 x 8 mm), either of tanned collagen or polytetrafluoroethylene, that were sutured in the infrarenal inferior vena cava of pigs and removed 1 hour after implantation. There was no significant difference between the extent of the thrombus-lined graft surface in the biologic group and that in the polytetrafluoroethylene group. The amount of inner thrombus on tanned collagen grafts was significantly correlated to platelet activity. Long-term experiments involved 30 similar segments of both materials, which were sutured in the inferior vena cava and harvested 7, 14, 28, 56, and 112 days after operation. The 112-day patency rate of collagen grafts was 67%. The 56-day patency rate of polytetrafluoroethylene grafts was 16%. The difference was statistically significant (p less than 0.01). Collagen grafts were lined by a thin neointima (200 micron) in all but two cases. The neointima was completely endothelialized within 4 weeks from implantation. In conclusion, tanned collagen grafts may represent a suitable material for venous replacement.     

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 carotidfemoral 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.     

Improved spontaneous endothelialization by postfixation treatment of bovine pericardium

The longstanding release of locally cytotoxic glutaraldehyde concentrations from fixed biological materials is one reason for the lack of spontaneous endothelialization in vivo and in vitro of biological grafts (and valves). Preliminary studies have shown that bovine pericardium postfixation-treated with aminoacids may be endothelialized in vitro. To test whether such treatment improves spontaneous endothelialization in vivo 8 cm long segments grafts (6 mm I.D.) were interposed into the carotid arteries of 10 sheep. Ten grafts were sewn from postfixation treated pericardium (PTP), 4 from standard glutaraldehyde fixed pericardium (GAP) and 6 polytetrafluoroethylane (PTFE) grafts were implanted to study the degree of spontaneous endothelialization. In two pregnant animals, all grafts occluded (2 PTP, 1 GAP, 1 PTFE). In the remaining animals 1 additional PTFE graft occluded and 2 PTFE and 1 GA grafts showed significant thrombotic obstruction. The patent grafts were harvested at 3 months and the surface covered with red thrombus was determined by planimetry. PTP grafts had significantly less (12.2% +/- 15%) thrombotic appositions than glutaraldehyde treated grafts (49% +/- 20%; P = 0.01) and PTFE grafts (40.5% +/- 13%; P = 0.01). In the central areas of the PTP grafts, endothelial cells spread directly on the collagenous matrix and produced a basal lamina. In GAP, endothelial cells spread on amorphous material or macrophages and in PTFE grafts, only a neointima composed of myofibroblasts was endothelialized. Neutralization of glutaraldehyde concentrations by postfixation with aminoacids improves spontaneous endothelialization in vivo in biological materials.     

骨髓CD34+细胞人工血管内皮化实验研究

目的探讨骨髓CD34^ 细胞种植的人工血管内皮化和通畅性。方法取犬骨髓,经免疫磁珠分离出CD34^ 细胞,内皮细胞生长因子(vascular endothelial grow factor,VEGF)诱导并扩增;种植聚四氟乙烯(polytetra fluoro ethylene,PTFE)人工血管,再植入犬的下腔静脉和腹主动脉。结果骨髓分离的细胞经流式细胞仪鉴定为CD34^ 细胞,CD34^ 细胞经VEGF诱导培养后经免疫细胞化学和透射电镜检测证实为内皮细胞。内皮化人工血管植入动脉闭塞率0,狭窄率12．5％;人工血管植入静脉闭塞率12．5％,狭窄率25．0％。结论骨髓经免疫磁珠可分离出CD34^ 细胞,CD34^% 细胞经VEGF诱导可定向分化为内皮细胞。骨髓CD34^ 细胞经体外培养种植于PTFE人工血管,获得理想的内皮化和通畅率。     

Surface morphology of circulating platelets: a suggested parameter for the monitoring of endothelial seeded grafts

For the clinical follow-up of endothelial cell seeded grafts, non-invasive but sensitive parameters are required in order to determine the extent and effectiveness of the endothelialization process. Such a very sensitive parameter was found by investigating the surface morphology of circulating platelets. This technique is based on the previous observation that characteristic alterations of platelet morphology are closely associated with the course of the aggregometer tracings. Applying these ultrastructural criteria we investigated 16 patients after distal femoro-popliteal and femoro-crural bypass surgery. Reinforced 6 mm non-seeded PTFE vascular grafts were compared with reversed autologous saphenous vein grafts. As early as 4 hours after graft implantation, a significant higher percentage of discoid platelets with marginal pseudopods--representing the earliest stage of platelet activation--as well as of spheroid "shape changed" platelets was found in patients with PTFE grafts. Moreover, perforated platelet membranes were exclusively found in the PTFE group. These findings were most distinguished 10 days postoperatively but persisted over the follow-up period of 1 year. Considering the sensitivity of the described technique it might provide a simple, reliable, and non-invasive means for the clinical assessment of successful graft endothelialization.     

The Vicious Cycle of Nonhealing Neointima in Fabric Vascular Prostheses

Abstract: Delayed neointimal healing of a fabric vascular prosthesis was investigated in an animal study focusing on the relationship between red thrombus, fibrinolysis, and endothelialization on the luminal surface. Fabric vascular prostheses were implanted into the descending aortas of 72 dogs. Fifty-nine grafts were explanted from 1 h to 1,705 days after implantation. One hour after implantation, the graft wall was red in color due to fresh thrombus; however, at 1 day the luminal surface became white. Red thrombus reappeared at 1 week and remained present in the long-term. Microscopically the initial red thrombus contained numerous erythrocytes. The white thrombus at 1 day was composed of a dense fibrin network without erythrocytes. At 2 days numerous lacunae appeared in the fibrin layer, and at 3–5 days cavernae and low density fibrin areas were present secondary to fibrinolysis. These areas allowed the blood components to infiltrate into the fibrin layer, and as a result red thrombus reformed within it. The thrombi on the luminal surface in the long-term was always red in color and composed of complicated, multiple stages of thrombus formation, i.e., fresh thrombus with erythrocytes, dense fibrin without erythrocytes, low fibrin density areas, lacunae and cavernae in the fibrin layer, and blood component infiltration into these spaces. Thrombus was always newly formed and present, and involuted in parallel due to fibrinolysis, suggesting that these phenomena perpetuated in a vicious cycle. However, at the anastomoses fibrinolysis was present, but blood component infiltration was prevented by the endothelial cell lining. These results suggest that endothelialization may arrest the vicious cycle of non-healing neointima in fabric vascular prostheses.     

Expanded polytetrafluoroethylene as dialysis access grafts: serial study of histology and fibrinolytic activity.

Expanded polytetrafluoroethylene (Gore-Tex) grafts were evaluated as hemodialysis access shunts. A serial study of cellular ingrowth and endothelialization was performed. The development of fibrinolytic activator was sequentially monitored and shown to be consistently higher in the Gore-Tex grafts than in either knitted or woven Dacron prostheses. Functional assessment by repeated needle puncture demonstrated that PTFE grafts were durable, had uniform patency (6 mm ID size) and showed no evidence of degeneration during the time limit of this study. Although these are favorable characteristics, further long-term evaluation of PTFE grafts is necessary before its role as a vascular prosthesis applicable to clinical practice, can be established.     

Morphologic quantitation of pseudointima and effects of antiplatelet drugs on vascular prostheses in goats

The natural process of endothelialization, pseudointimal formation, and connective tissue incorporation of the expanded PTFE grafts in the goat was documented through histologic examination of specimens harvested at 2, 4, 6, and 8 weeks. The goats demonstrated a progressive increase in pseudointimal pannus ingrowth from the anastomoses at a rate of 11.3 mm over a 12 week period. Histologic changes according to time of vascular graft incorporation in the goat model were found to be comparable to those of the dog, pig, and calf models reported in the literature. Platelet-inhibiting drugs, aspirin, dipyridamole, nifedipine, and ibuprofen were administered to goats after replacement of their infrarenal aorta with 5 cm of 8 mm diameter expanded PTFE grafts. The effects of the drugs on graft endothelialization and anastomotic pseudointimal formation was compared with those in the untreated control group after 12 weeks. Aspirin and dipyridamole had no detrimental effect on the healing process compared with the untreated control group. Studies with nifedipine and ibuprofen did not demonstrate a decrease in pseudointimal hyperplasia. Antiplatelet treatment resulted in no significant change in the rate of endothelialization of expanded PTFE grafts.     

Comparison of the deployment and healing of thin-walled expanded PTFE stented grafts and covered stents

This study evaluated the deployment and short-term healing of thin-walled expanded polytetrafluoroethylene (ePTFE) stented grafts and covered stents as endoluminal prostheses in normal canine aortas and in an abdominal aortic aneurysm (AAA) model. Stented grafts consisted of a 7 cm length of 3 mm internal diameter ePTFE graft (Impra, Inc., Tempe, Ariz.) with two P-188 Palmaz stents (Johnson & Johnson Interventional Systems, Warren, N.J.) deployed along the inner surface of the ends of the graft to secure the prosthesis to the arterial wall. Covered stents were fabricated by placing a 3.7 cm length of 3 mm internal diameter thin-walled ePTFE graft over a P-394 Palmaz stent. Four covered stents and four stented graft prostheses (two of each prototype in the normal canine aorta and AAA model) were implanted in eight animals. One prosthesis of each type in each model was removed at 30 days and one at 60 days. Prior to removal, prostheses were evaluated by CT scan, arteriography, and intravascular ultrasound imaging with values compared to those obtained when the prostheses were deployed. Gross inspection and microscopic evaluation were performed at scheduled explantation. In general, the stented grafts were more difficult to accurately deploy. Healing and maintenance of long-term patency without significant luminal obstruction or occlusion occurred in only one 30-day sample in a normal canine aorta. The 30-day stented graft specimen that had been implanted in an AAA and required the addition of a covered stent to seal a maldeployment of the distal segment was also patent. The 60-day stented graft in the normal canine aorta was occluded with narrowing of the graft between the stents. The 60-day stented graft in the AAA was patent with one central fold and thrombus occupying approximately 20% to 30% of the lumen at this site. In contrast, the covered stent devices were less difficult to accurately deploy. All of the covered stent devices were patent with well-incorporated surfaces. Deployment of covered stents was more accurate and less complicated compared to stented grafts. All patent endoluminal prostheses and stent surfaces were well incorporated into aortic tissues. Problems with graft narrowing, folding, and subsequent thrombosis occurred in the unstented segments of the stented grafts. These preliminary findings support the further development and use of completely supported (stented) devices as endoluminal prostheses.Supported by Impra, Inc., Tempe, Ariz.Presented at the Fourteenth Annual Meeting of the Southern California Vascular Surgical Society, September 15–17, 1995, La Jolla, Calif.     

Experimental studies on application of small-caliber vascular prosthesis produced by polyurethane.

It has been suggested that a microporous structure enhances fast and complete endothelialization. For long-term patency, antithrombogenicity and microporous structure are very important factors. In this paper, we have developed a new technique to give a micro-porous structure to small-caliber vascular prosthesis produced by polyurethane which has favorable antithrombogenecity. A mixed solution (tetrahydrofuran: dimethylformamide = 1:1) containing 13 wt% of segmented polyurethane and a variable amount of calcium carbonate (mean particle size of 8 mm in diameter) was dip-coated on a glass mandril of 3 mm and 6 mm in diameter and placed into distilled water for 24 hours. After the glass mandrill was removed, this polyurethane tube was placed into 1 mmol hydrochloric acid for 1 hour, and a microporous polyurethane vascular prosthesis of 20 mm in length was completed. These prostheses of 3 mm and 6 mm in diameter were implanted into the femoral and the carotid arteries, and the abdominal aorta of the dogs, respectively. Patency was recognized by arteriography and Duplex scanning and the removed grafts were inspected macro- and microscopically. Greater hydraulic permeability of this graft was obtained with an increase in the quantity of calcium carbonate mixed with polyurethane. In elasticity, this graft was more similar to the canine jugular vein than the polytetrafluoroethylene graft. Patency was observed 8 weeks after implantation on the arteriogram, and neointima was observed microscopically on the smooth and lustrous lumen. The new polyurethane vascular prosthesis we developed might provide a potential prosthesis for small-caliber vascular reconstruction.