J
THoRAc CARDIOVASC SURG
1991;102:440-7
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 polytetraftuoroethylene, 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 polytetraftuoroethylene 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 1l2-day patency rate of collagen grafts was 67 %. The 56-day patency rate of polytetraftuoroethylene grafts was 16 %. The difference was statistically significant (p < 0.01). Collagen grafts were lined by a thin neointima (200 #Lm) 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.
G. B. Ratto,a P. Romano," M. Truini," M. Fraseio," S. Rovida," A. Badini,c and D. Zaccheo," Genoa, Italy
Eency rates of synthetic and biologic grafts in the venous system often have been disappointing, 1-10 presumably as a result of either the unfavorable hemodynamic conditions in veins or the thrombogenic characteristics of currently available prosthetic materials.l" II At present, the only suitable venous substitute is autogenous vein. Large veins have been replaced successfully with spiral grafts constructed from autogenous saphenous veins.12- 14 However, the technique is laborious and requires an extensive great saphenous vein excision." For grafts with low blood flow rates a thrombogenic surface is an unacceptable property, leading to luminal thrombus accumulation and early failure. Consequently, several concepts have been developed to improve blood compatibility of graft surfaces. They include (1) modification of existing prosthetic materials by either heparin bonding'- or endoFrom the Institutes of Clinica Chirurgica I,a Anatomia Umana Norrnale.Pthe Departments of Anatomia Patologica II1,e Semeiotica Chirurgica H,d and Statistica Medica; University of Genoa, Genoa, Italy. Received for publication July 19, 1989. Accepted for publication April 4, 1990. Address for reprints: Professor Giovanni B. Ratto, Istituto di Clinica Chirurgica la Universitadi Genova,Viale BenedettoXV, 10, 16132 Genova, Italy.
12/1/21575
440
thelial seeding 16,17 and (2) the search for new vascular structures, such as the polyglactin 910 suture mesh graft 18 and the modified mandril-grown grafts. 19, 20 '. The present research study was undertaken to evaluate the potential of glutaraldehyde-tanned mandril-grown vascular grafts as caval substitutes. Materials and methods The tanned ovine collagen prosthesis, a composite graft of ovine collagen (mainly types I and III) with a polyester mesh skeleton, 19,20 has been tested. As control material, the expanded polytetrafluoroethylene (ePTFE) ringed vascular graft, representing the standard for venous replacement in humans, has been used. Experimental design. The experimental plan included shortterm and long-term experiments (Fig. 1). Short-term experiments were performed to assess early phenomena characterizing blood-prosthesis interaction. Thirty tubular grafts of either tanned ovine collagen or ePTFE were implanted in the infrarenal inferior vena cava (IYC) of pigs and removed 1 hour after circulation had been restored through the implant. Long-term experiments were performed to evaluate late phenomena leading to graft healing or failure. Thirty prostheses of both materials were used as lye substitutes and harvested at scheduled intervals of 7, 14, 28, 56, and 112 days after operation. The infrarenal lye was chosen as the implantation site, because it has been demonstrated to represent the ultimate test of any venous substitute.': 21 To assess potential causes of early or late thrombosis of ovine collagen grafts, we correlated a series of animal-related variables with short-term and long-term results.
Volume 102 Number 3 September 1991
Venous substitutes
-+
C A
H
.:
THROMBUS LINED:3
T:O
o
BIOLOGIC
GRAFTS: 15
-,THROMBUS FREE:12
441
BIOLOGIC 15 GRAFTS:
R:2
+ T:O
R:2
U:13
C
1
T:2
""--t.,, ""--t.,, n
u:11i:2
T:l
T,O
x
X
P
P
u:o
E THROMBUS LINEO:4
/'
r1
E-PTFE GRAFTS: 15
E N
-,
T:S E-PTFE GRAFTS: IS-+R:2I'S
M
E
THROMBUS FREE:11
U:8~R:2
U:l-+R:O T
S _____ IH ~EGEND:
T:O U:1------------
S /
T:O U:O
/
T = THROMBOSED GRAFTS
7__14
28
= REMOVED PATENT GRAFTS
56
/
/ __ 112_ _ DAYS
U = UNREMOVED GRAFTS
Fig. 1. Flow diagram illustrating the experimental design and the relevant results. The following parameters were investigated: hematocrit value, plasma fibrinogen level, platelet number and aggregation, and plasma and blood viscosity. Blood samples (10 ml) were withdrawn from the cannulated external jugular vein before starting operation and 1 hour after blood-prosthesis contact. Animal models. Thirty pigs weighing 20 to 23 kg were used for both short-term and long-term experiments. The animals were randomly assigned to receive tanned ovine collagen or ePTFE grafts. They were premedicated with ketamine (7 tug] kg, intramuscularly). The animals were intubated and ventilated with nitrous oxide and oxygen (5:4) with a mechanical ventilator. Anesthesia was maintained with enflurane (Ethrane). Operations were performed under strict sterile conditions. No antibiotics and no systemic heparinization were used. Intravenous infusion of Ringer's lactate solution was set up at the rate of 300 mljhr in every instance. Through a midline incision the IYC was exposed from the renal veins to the bifurcation. The infrarenal IYC was excised, leaving adequate cuffs for subsequent anastomoses, and replaced with 3.5 ern long (diameter 8 mm) prosthetic segments. Before implantation tanned ovine collagen grafts were rinsed twice with 20 ml of saline solution (containing 10,000 units of heparin per liter). The graft was clamped at one extremity and filled with 50,000 units of heparin. The other end of the graft was also clamped, and the prosthetic segment was placed for 10 minutes in a basin containing saline solution. Clamps were then removed, and heparin was allowed to drain out of the graft. Polypropylene (6-0) running sutures were used for end-to-end anastomoses. Care was taken not to narrow the anastomoses. No significant disparity existed
between the graft and the IYC diameter. One hour after implantation the grafts were removed, rinsed in saline solution, longitudinally cut open, fixed on a board, grossly inspected, and photographed. The thrombus-free surface area was calculated by the grid rnethod.P For analysis, exclusively the middle portion of each graft was considered. The 1 em graft portion from each suture line was cut away, because hemodynamic factors might have accounted for local thrombus deposition. Specimens were classified into three categories: (I) the thrombus-free group, in which thrombus deposition covered less than 10% of the graft flow surface; (2) the thrombus-lined group, in which more than 50% of the flowsurface was lined with thrombus; and (3) the indeterminate group, in which thrombus deposition covered more than 10% but less than 50% of the flow surface. Specimens falling into the indeterminate category were discarded, and the experiments were repeated with other animals. All excised grafts were examined by light (LM) and scanning electron microscopy (SEM). Identical prosthetic segments (35 X 8 mm), which were intended for the evaluation of long-term graft performance, were then implanted to bridge the gap in the IYC, according to the aforedescribed technique. Implant patency was assessed 7 and 14 days after operation by means of a postoperative phlebogram and/or at the time of graft removal. As soon as failure was detected or animals appeared in poor health conditions, prosthetic segments were excised. At the completion of the observation period animals were anesthetized and given 10,000 units of heparin intravenously. The grafts were removed intact with a small segment of adjacent vena cava at each anastomo-
The Journal of Thoracic and Cardiovascular
4 4 2 Ratto et al.
Surgery
Table I. Tanned ovine collagen grafts in the IVe in short-term experiments; effects of host-related factors on blood-prosthesis interaction Group" Factor
Hematocrit (%)) Fibrinogen (mg/dl) Platelets (no. X 103/mm 3)t Platelet aggregation (%):j: (ADP) Maximal After 5 min Platelet aggregation (%) (collagen) Maximal After 5 min Platelet aggregation (%) (epinephrine) Maximal After 5 min Plasma viscosity (160 sec") (cP) Blood viscosity (cP) (shear rate) 40 sec -I 120 sec"! 160 sec"!
1 (12 cases) Mean ± SD
2 (3 cases) Mean ± SD
29.5 ± 2.99 403 ± 220 421 ± 107
29.3 ± 2.51 329 ± 133 659 ± 179
45.0 ± 12.8 22.7 ± 15.9
65.3 ± 14.1 45.6±5.13
34.6 ± 23.9 31.1 ± 22.5
45.3 ± 8.38 39.3 ± 13.6
3.66 ± 2.49 3.25 ± 2.41 1.49 ± 0.11
3.33 ± 0.57 3.33 ± 0.57 1.42 ± 0.06
3.89 ± 0.32 3.22 ± 0.27 3.04 ± 0.20
3.43 ± 0.57 2.90 ± 0.52 2.80 ± 0.53
SO. Standard deviation; A DP. adenosine diphosphate. 'Group I represents thrombus-free grafts; group 2 represents thrombus lined grafts. tp < 0.01. :j:p < 0.05.
sis, rinsed in saline solution, longitudinally opened, and examined by LM and SEM. Cultures of all harvested specimens were obtained. Preparation for LM and SEM. The portions for LM were fixed in 10% buffered formaldehyde solution and embedded in glycol methacrylate. Semithin sections (0.5 to 1.0 JLm) were obtained and stained with hematoxylin-eosin, Weigert's, Gram's, and van Gieson's stains. The neointima thickness and the number of mononuclear cells (number/100 JLm2) infiltrating the graft wall were measured with a calibrated square grid (292 JLm broad), according to the method described by Cengiz and colleagues.P For every graft there were 10 measurements on each of the six sections cut in a staggered manner. Mean values were obtained by averaging the values in each section and then by averaging the six sections together. No correction for cellular overestimation was made because this was a comparative study in which absolute values were not prominent. Grafts portions that were intended for SEM were fixed in cold 2% buffered glutaraldehyde solution (pH 7.4). Specimens were left in the fixative for 2 hours at 4 0 C, postfixed in 1% osmium tetroxide, dehydrated in a graded series of ethanols, and dried by the critical point method. Specimens were then sputter-coated with gold palladium and examined with an lSI S5 40 scanning electron microscope (International Scientific Instruments, Inc., Milpitas, Calif.).
Clotting and rheologic profiles. The hematocrit value was determined according to the microhematocrit method. Fibrinogen levelswere evaluated by means of the functional method.P The tendency of platelets to aggregate was assessed according to the turbidimetric method by adding adenosine diphosphate at different concentrations (5, 2.5, 1,0.5 JLmol/L),epinephrine (10 JLmoIjL), and collagen (4 JLg).25 Both maximal and 5 mm platelet aggregation were calculated and expressed as percent of transmittance. Plasma and blood viscosity were measured with a Haake CV 100 viscosimeter (Haake/Fisons Instruments, Valencia, Calif). Plasma viscosity was determined at usual temperature (37 0 C) and shear rate (160 sec/L). Blood viscosity was measured at 37" C under various shear rates (40, 120, 160 sec/L). The experimental error with the method used was calculated and found to be about 1%. Blood viscosity was adjusted to 30% hematocrit value with a semilogarithmic graph of viscosity plotted against hematocrit. Statistical analysis. Each morphologic evaluation was done independently by two pathologists in a blinded manner. The results have been considered when there was concordance between the two observers. Cumulative patency rates were calculated by the actuarial method of Kaplan and Meier, according to the type of graft material used. Patency curves were compared by the Mantel-Haenszel method. Results of the investigated coagulative and rheologic factors were expressed as mean ± standard deviation. Differences in mean values between graft groups were assessed with the Kruskal-Wallis one-way analysis of variance by ranks and the Mann-Whitney U test. Results LM examination revealed that before implantation the biologic grafts have a thin collagen tissue cover over the Dacron mesh. SEM revealed that this cover' presents small tears, exposing synthetic fibers. The silver staining method did not support the observation of exposed mesh bundles, suggesting that the noted small tears originated during processing for SEM examination. Short-term experiments. In the ovine collagen group 12 grafts were classified as thrombus-free and three as thrombus-lined. One graft had fallen into the indeterminate category. In the ePTFE group 11 segments were thrombus-free and four thrombus-lined. Three experiments were repeated because of indeterminate results. There were no significant differences between the thrombus-free surface scores in the ovine collagen group and those in the ePTFEgroup. With regard to preoperatively measured coagulative and rheologic factors, animals receiving ovine collagen grafts were matched with those animals undergoing ePTFE graft implantation. Table I shows the relationship between measured host-related factors and the amounts of inner thrombus deposited on ovine collagen grafts. A significant correlation was found between either platelet number (p < 0.01) or function (p < 0.05) and the extent of thrombus-lined inner surface. One hour after implantation the platelet number was consistently reduced in the thrombus-lined collagen
Volume 102 Number 3 September 1991
Venous substitutes
443
Fig. 2. LM and SEM appearance of an ovine collagen graft 1 hour after implantation; a single mononuclear cell layer covers the graft flow surface. (A, Hematoxylin-eosin, magnification X25. B, SEM X6000.)
graft group but not in the thrombus-free group (p < 0.05). A postoperative decrease in platelet function wasfoundin both groups, but the phenomenon wasmore severe in the thrombus-lined group than in the thrombusfree group (p < 0.05). The flow surface of the removed ovine collagen grafts classified as thrombus-free showed different morphologic patterns. In four cases the graft surfaceunderwentno demonstrable changes, remaining in stable equilibrium with the circulating blood. In five cases a singlelayer of mononuclear cellshomogeneously covered the innergraft components (Fig. 2). In three cases smallamounts of fibrin and platelets were exclusively seen where Dacron fibers appeared on the graft surface
(Fig. 3); the remaining blood-prosthesis interface was lined with a singlemononuclear cell layer. Long-term experiments. Patency rates of both biologic and synthetic grafts are shown in Fig. 4. Overall patency for ovine collagen grafts was significantly superior to that for ePTFE grafts (p < 0.01). In the biologic graft group two animals appeared sick and were thereby killed 28and 55daysfromimplantation,respectively. The grafts were found to be occluded. The relationship between long-term ovine collagen graft patency and host-related factorsisshownin Table II. Platelet function (p < 0.01), plasmafibrinogen levels (p < 0.05), and plasma viscosity (p < 0.05) were significantly higher in ani-
4 4 4 Ratto et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 3. LM appearance of an ovine collagen graft I hour after implantation; thrombus deposition is limited to the exposed Dacron fibers-. (Hematoxylin-eosin, magnification X I0.)
malswith occluded grafts than in thosewithpatent grafts. The flow surface of both collagen and ePTFE grafts was initially(7 to 14days after implantation) covered with a network of fibrin fibers, enmeshing platelets and blood cells. From 28 days after implantationthe flow surfaceof patent biologic grafts was completely and regularly endothelialized. The endothelial covering was supported by a layer of vascularized mesenchymal tissue, showing no evidence of degenerative phenomena. In this newly formed connective layer, elastic fibers were neveridentified. During the entire observation period the neointima was loosely attached to the prostheticwall and detached s\JOntaneousl)' as the collagengrafts were cut open.Consistent variationsof the thickness of the neointimallayer were found among different biologic prostheses (Fig. 5).
In \()cases tbenoointima tbicKness was invariab\)' inferior to 200 ~m, whereasin two cases (removed 28 and 56 days after implantation,respectively) it ranged from 500 to 700 ~m. Platelet function was the only host-related measurementthat provedto be associated with the development of a thick neointimal layer (p < 0.01). Patent prostheticsegmentsthat developed a thin neointimahad a significantly greater reducedinfiltration by mononuclear cells (9.1 ± 3.6/100 ~m2) than both occluded grafts (19.7 ± 2.5/100 ~m2) (p < 0.001) and patent grafts lined with a thick neointima (31.7 ± 2.2/100 ~m2) (p < 0.001). The neointima that developed on the only long-term functioning ePTFE graft was uniformly thinner than 200 ~m and was completely linedwith an endothelium-like layer.
Discussion The lack of readly available, currently successful venous substitutes makes the search for new prosthetic materials pressing. The present study has indicated that a conduitcomposed of tanned ovine collagen and polyester mesh may fulfill major requirements for an ideal venous graft, including early and complete endothelialization of its flow surface. Previous experimental studies demonstrated complete healing of ovine collagen grafts implanted in the arterial system. 19, 20 The present study has shownthat completehealingalsomay be achieved in the venous system. This is an important issue, because earlierresearchfrom ourlaboratoryrevealedthat healing of venous grafts is generally prevented or delayed, with respect to healing of arterial grafts, because of larger inner thrombus deposition.26 'Thrombus deposition onthe flow surfaceof ovine collagen grafts in the Iv'C wasminimal,favoring the development of a thin neointima. Only two prostheses had a neointima thicker than 500 urn. Possible causesfor thisunfavorable behavior couldbe (1) an enhancedplatelet activityor (2) a persistent immunologicresponse to tanned ovinecollagen tissue, as indicated by the moreintensemononuclear cellinfiltration of the prosthetic wall. Unfortunately, no further characterization of these infiltrates (i.e., immunohistochemical analysis of mononuclear cell subsets) could be attempted, whichdependedon the paucityof reagentsin the pigspecies. The immunologic hypothesis might be supported by (I) the delayedoccurrenceof collagengraft failures, possiblyrelated to the procoagulantactivityexertedby infil-
Volume 102 Number 3 September 1991
Venous substitutes
445
PATENCY RATES
r
100
90
•
•
I TANNED OVINE COLLAGEN GRAFTS
•
•
I PTFE GRAFTS
I
I
80 70
60
50 40 30
20
•
...- - - - - . . - - - - f r l - - - -•
10
o.
L---r---,---,..---,.----{
7
14
21
28
,1-'----,r------.r---I
56
112
DAYS
Fig. 4. Patency rates of tanned ovine collagen and ePTFE grafts in the lye.
trating mononuclear cells, and (2) results from previous studies with differentexperimental models, showing the maintenance of someantigenicactivityin ovine collagen grafts.I 9 Because biologic conduits represent a major direction in the development of vascular substitutes, 15.27-29 one might ask whether and why ovine collagen prostheses would be superior to previous fibrocollagenous conduits. Theoretic reasons for this superiority could be (1) better intermolecular cross-linking by glutaraldehyde tanning; (2) the incorporated, rather than externally applied, polyester mesh, reducing the graft rigidityassociated with the periprosthetic response to synthetic materials; (3) improved structural properties by sophistications in graft manufacturing (acting on the design and tension of the polyester mesh over the mandril, the thickness and homogeneity of the graft inner collagen layer can be modified). Dilation and aneurysm formation are the major causes of collagenous tube failure in the arterial systemr" the low mechanical forces characterizing venous circulation also might have accounted for the excellent viability of our collagen grafts. The fewarticles dealing with the use of fibrocollagenous conduits in the venous system supportthe soundness of biologic grafts as venous substitutes. Horsch and colleagues'? implanted modified bovine carotid artery heterograftsin the iliocav-
al tract of dogsand obtaineda patency rate of87.5% 210 daysafter operation. Sitzmann and colleagues- J reported that cryopreserved venous allografts interposed in the venous systemuniformly occluded within 5 weeks. However,an 83%early patency rate was achieved in animals treated with antiplatelet drugs. Moreover, our findings suggestthat the characteristics of the presentlyavailable ovine collagen grafts could be further enhanced. The connective tissuelayercovering the syntheticmeshis thin and easilydamaged by handling and manipulation during operation. The fact that damages in the inner collagen layer may induce local thrombosis has been demonstrated by the observation of fibrin deposits confined to areas corresponding to surface imperfections. To overcomethis problemnewtypesof ovine collagenprostheses with a thicker inner fibrous layer are being evaluated. This study alsoshowed that the treated ovinecollagen grafts havea lowtendency to activateplateletaggregation and innerthrombusformation. Blood compatibility of the graft surface might be surprising, because collagen is known to be a potent plateletaggregator.A recent experimental study'? showed that coating the flow surface of synthetic grafts with bovine collagen does not enhance plateletaggregation. Intermolecularcross-linking during graft fabrication was proposed as the mechanism inhibiting the proaggregating activityof collagen. Conversely,
The Journal of Thoracic and Cardiovascular Surgery
4 4 6 Ratto et al.
Table II. Tanned ovine collagen grafts in the It/C in long-term experiments; effects of host-related factors on long-term patency of Il/C substitutes Group* Factor
Hematocrit (%») Fibrinogen (rng/dlj] Platelets (no. X 1()3/mm 3) Platelet aggregation (%)t (ADP) Maximal After 5 min Platelet aggregation (%)t (collagen) Maximal After 5 min Platelet aggregation (%):j: (epinephrine) Maximal After 5 min Plasma viscosity (cP)t Bloodviscosity (cP) (shear rate) 40 sec"! 120 sec- 1 160 sec- 1
Fig. S. Gross appearance of two ovine collagen grafts 28 days after implantation. A, Thin neointima graft «200 ,urn). B, Thick neointima graft (>500 ,urn).
platelet activity was the only host-related factor significantlycorrelatedto eithershort-term (luminal thrombus deposition) or long-term (graftperformance) results. This would suggest that drug control of platelet aggregation mightfurther reducethrombus formation on ovine collagensurfaces and improve long-term graft patency. Graft patencywasalsocorrelated to plasmafibrinogen concentrations and viscosity values. Accordingly, previous reportsemphasized the majorroleoffibrinogen inthromboticocclusion ofgraftspositioned inthevenous system-? We are aware that somecriticisms couldbe raisedto our experimental model. It couldbe disputed that results fromlaboratoryanimalscannotbemerely extrapolated to humansin consideration of the great differences in graft healing characteristics. The speed and completeness of the healing process in the growing pig highly differentiates it from humans.P However, the results we have obtained maintain a clinical relevance, because (I) the development of the neointima is basically the same process in humans and experimental animals'" and (2) the growing pig tends to develop a thick and unstable inner capsule, makingit an idealmodel for assessing the ability ofanytestedgraft materialto keepthe neointima thinand viable. 33 A second criticism mightconcern theinadequate length of the prosthetic segments." Nevertheless, in the venous system even short prosthetic segments are at high risk for failure," as confirmed by the elevated occlusion rate obtainedwith externally stentedePTFE prostheses, despite the fact that ePTFE wasshown previously to yield
I (12 cases) Mean ± SD
2 (3 cases) Mean ± SD
30.1 ± 2.12 337 ± 120 448 ± 161
27.0 ± 4.035 594 ± 362 551 ± 85.1
47.9 ± 16.8 23.1 ± 16.6
53.6 ± 3.21 44.0 ± 3.60
31.9 ± 21.8 27.5 ± 20.1
56.3 ± 3.21 54.0 ± 3.60
2.83 ± 1.11 2.50 ± 1.00 1.45 ± 0.07
6.66 ± 3.21 6.33 ± 3.05 1.60 ± 0.14
3.78 ± 0.44 3.15 ± 0.36 2.97 ± 0.29
3.88 ± 0.18 3.17 ± 0.29 3.07 ± 0.30
SD. Standard deviation; ADP. adenosine diphosphate. ·Group 1 represents patent grafts; group 2 represents occluded grafts.
tp <0.05. < om.
:j:p
patency rates better than those of other available graft materials.'?
In conclusion, ovine collagen grafts may represent a reliable prosthetic material for venous replacement. The administration of antiplatelet drugs immediately on restoration of blood flow should have beneficial effects on graft patency. Further refinements in graft manufacture might improve its long-term performance. The assistance of M. Galanti during the experiments and in the preparation of this article is gratefully acknowledged. REFERENCES 1. Haimovici H, Hoffert PW, Zinicola N, Steinman C. An experimental and clinical evaluation of grafts in the venous system. Surg Gynecol Obstet 1970;31:1173-86. 2. Scherck JP, Kerstein MD, Stansel HC. The current status of vena cava replacement. Surgery 1974;76:209-33. 3. Calvert MH, Gough MH. The effects of anti-thrombotic drugs on prosthetic grafts in the canine inferior vena cava. Br J Surg 1976;63:910-3. 4. Fiore AC, Brown JW, Cromatic RS, et al. Prosthetic replacement for the thoracic vena cava: an experimental study. J THORAC CARDIOVASC SURG 1982;84:560-8.
Volume 102
Number 3 September 1991
5. Robison RJ, Peigh PS, Fiore AC, et al. Venousprostheses: improved patency with external stents. J Surg Res 1984; 36:306-11. 6. Plate G, Hollier LM, Gloviczki P, Dewanjee MK, Kaye MP. Overcoming failure of venous vascular prostheses. Surgery 1984;96:503-10. 7. Gloviczki P, Hollier LH, Dewanjee MK, Trastek VF, Hoffman EA, Kaye MP. Experimental replacement of the inferiorvena cava:factors affectingpatency.Surgery 1984; 95:657-66. 8. Yamamoto N, Takaba T, Hori G, et al. Reconstruction with insertion of expanded polytetrafluoroethylene (EPTFE) for iliac venous obstruction. J Cardiovasc Surg 1986;27:697-702. 9. Dartevelle P, Chapelier A, Navajas M, et al. Replacement of the superior vena cava with polytetrafluoroethylene grafts combined with resection of mediastinal pulmonary malignant tumors. J THORAC CARDIOVASC SURG 1987; 94:361-6. 10. Ribbe EB, JonssonBA, Norgren LEH, Strand SE, Thorne JL. Platelet aggregation on peritonealtube grafts and double velour grafts in the inferior vena cava of the pig. Br J Surg 1988;75:81-5. II. Plate G, Hollier LH, Fowl RJ, Sande BA, Kaye MP. Endothelial seeding of venous prostheses. Surgery 1984; 96:929-36. 12. Chiu CJ, Terzis J, MacRae ML. Replacement of superior venacava with the spiral compositeveingraft. Ann Thorac Surg 1974;17:555-60. 13. Doty DB. Bypassof superior vena cava: six years' experience with spiral vein graft for obstruction of superior vena cavadue to benignand malignant disease.J THORAC CARDIOVASC SURG 1982;83:326-38. 14. AndersonRP, Li Wei-I. Segmental replacement of superior venacava with spiral veingraft. Ann Thorac Surg 1983; 36:85-8. 15. Esquivel CO, Blaisdell FW. Why small caliber vascular grafts fail:a review of clinicaland experimentalexperience and the significance of the interaction of bloodat the interface. J Surg Res 1986;41:1-15. 16. Clarke JMF, Pittilo RM, Nicholson LJ, WoolfN, Marston A. Seeding Dacron arterial prostheses with peritoneal mesothelial cells: a preliminary morphological study. Br J Surg 1984;71 :492-4. 17. Nicolaides AN. Haemodynamicand rheological aspectsof vascular grafts. Acta Chir Scand 1987;538:12-7. 18. Bowald S, Busch C, Eriksson I. Arterial regeneration following polyglactin 910 suture mesh grafting. Surgery 1979;86:722-33. 19. Christie B, Ketharanathan V, Perloff LJ. Patency rates of minutevascularreplacements: the glutaraldehyde modified mandril-grown conduit. J Surg Res 1980;28:519-32. 20. Perloff LJ, Christie BA, Ketharanathan V, et al. A new replacement for small vessels. Surgery 1981;89:31-41. 21. Collins HA, Burrus G, DeBakey ME. Experimental eval-
Venous substitutes 447 uation of grafts in the canine inferiorvena cava. Am J Surg 1960;99:40-4. 22. Venkataramani ES, Senatore F, Feola M, Tran RM. Nonthrombogenicsmall-caliberhuman umbilicalveinvascular prosthesis. Surgery 1986;99:735-43. 23. Cengiz M, Sauvage LR, Berger K, et al. Effectsof compliance alteration on healing of a porous Dacron prosthesisin the thoracic aorta of the dog. Surg Gynecol Obstet 1984;158: 145-51. 24. Clauss VA. Gerinnungsphysiologische Schnellmethodezur Bestimmuning des Fibrinogens. Acta Haemat 1957;17: 237-46. 25. Born GVR. Aggregation of blood platelets by adenosine diphosphate and its reversal. Nature 1962;194:927-9. 26. Ratto GB, Di Primio R, Romano P, et al. New directions in the use of carbon as vascular graft material. Vase Surg 1988;22:244-52. 27. Sparks CH. Silicone mandril method of femoropopliteal artery bypass. Am J Surg 1972;144:244-9. 28. Sparks CH. Silicone mandril method for growing reinforced autogenous femoropopliteal artery grafts in situ. Ann Surg 1973;177:293-300. 29. Dardik H, Ibrahim 1M, Sussman B, et al. Biodegradation and aneurysm formation in umbilical vein grafts: observations and a realistic strategy. Ann Surg 1984;199:61-8. 30. Horsch S, Pichlmaier H, Walter P, Landes TH. Replacement of the inferiorvena cava and iliac veinswith heterologous grafts in animal tests. Surgery 1978;84:644-9. 31. Sitzmann JV, Imbembo AL, Ricotta JJ, McManama GP III, Hutchins GM. Dimethylsulfoxide-treated, cryopreservedvenousallografts in the arterial and venoussystems. Surgery 1983;95:154-9. 32. Harker LA, Slichter SJ, Sauvage LR. Platelet consumption by arterial prostheses: the effect of endothelializationand pharmacologic inhibition of platelet function. Ann Surg 1977;186:594-601. 33. Sauvage LR, Berger KE, Mansfield PB. Future directions in the development of arterial prostheses for small and mediumcaliber arteries.Surg Clin North Am 1974;54:21328. 34. Reichle FA, Stewart GJ, Essa N. A transmissionand scanning electronmicroscopic study ofluminal surfaces in Dacron and autogenousvein bypasses in man and dog. Surgery 1973;74:945-60. 35. Scott SM, Gaddy LR, Parra S. Pyrolytic carbon-coated vascular prostheses. J Surg Res 1980;29:395-405. 36. Candas B, BernierJ, Guidoin R. Thrombose aigue des protheses arterielles en Teflon microporeux: importance relative de la longueurde la greffeet des parametres thrombohematologiques. Innov Tech Bioi Med 1983;4:217-32. 37. Gloviczki P, Hollier LH, Dewanjee MK, Trastek VF, Kopesky KR, Kaye MP. Quantitative evaluation of ibuprofen treatment on thrombogenicityof expanded polytetrafluoroethylene vascular grafts. Surgery 1984;95:160-8.