Increased platelet deposition on polytetrafluoroethylene grafts after balloon catheter thrombectomy Fritz R. Bech, M D , Spencer W. Gait, M D , and Jack L. Cronenwett, M D ,
Hanover, N.H. Prosthetic graft rethrombosis after thrombectomy may be potentiated by increased thrombogenicity of the restored flow surface. This experiment compared platelet deposition on polytetrafluoroethylene (PTFE) grafts after balloon catheter thrombectomy with deposition on new, nonthrombosed grafts. Three models of graft thrombosis were studied in eight dogs with 4 mm diameter by 7 cm PTFE graft segments: (1) in vitro model: grafts filled with blood, stored in 37 ° C saline solution; (2) in vivo model: blood-filled grafts stored in subcutaneous tissue; and (3) in situ model: one end of grafts anastomosed to femoral or carotid artery as a blind tube. Duration of thrombosis (1, 2, and 3 weeks) was studied by initiating one graft of each type per week in each dog. After 3 weeks, nine thrombosed grafts per dog were harvested and graft thrombectomy was performed with a 3F balloon catheter. An ex vivo flow-controlled perfusion cir~it was then created in each dog and platelet deposition was measured during the initial 20 minutes of graft perfusion after ~11Inplatelet labeling. Thrombectomized grafts were compared with new, control grafts not previously exposed to blood, as well as with grafts exposed for I hour to blood or plasma. Compared with control grafts, platelet deposition was significantly increased on in vivo (3.7 times control; p < 0.01), in situ (2.6 times control; p < 0.05), and in vitro thrombosed grafts (2.0 times control; p < 0.05). Age of thrombus was not a significant source of variation. Blood or plasma exposure alone did not significantly increase platelet deposition. These data suggest that antiplatelet therapy may be important at the time of PTFE graft thrombectomy. (J VASC SURG 1990;7:804-11.)
Polytetrafluoroethylene (PTFE) grafts are used most often for infrainguinal arterial reconstruction in patients whose autogenous saphenous vein is unsatisfactory. Primary PTFE graft patency is inferior to that o f autogenous vein grafts, but limb salvage may be improved by reoperation to increase the secondary patency of failing or thrombosed PTFE grafts. 1,2 Although surgical thrombectomy is frequently used to salvage PTFE grafts, the outcome of these procedures, especially the early rethrombosis rate, has not been widely reported. Furthermore, po-
From the Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center. Supported in part by the NationalInstitutesof Health Grant R01HL35102. Presented at the SixteenthAnnualMeeting of the New England Societyfor VascularSurgery,BrettonWoods, N.H., Sept. 2122, 1989. Reprint requests: Jack L. Cronenwett,MD, Sectionof Vascular Surgery,Dartrnouth-HitchcockMedicalCenter,Hanover,NH 03756. 24/6/20069 804
tential changes in graft thrombogenicity after thrombectomy have not been studied. The purpose of this investigation was to develop an animal model to study graft thrombectomy and early reocclusion that would initially address two specific questions: (1) Are thrombosed PTFE gra~Ls, once thrombectomized, more thrombogenic than PTFE grafts not previously exposed to blood? and (2) Does the duration of graft thrombosis affect thrombogenicity after thrombectomy? Because platelet accumulation has been shown to have a major role in the early thrombosis o f newly implanted PTFE grafts, ~,4 this study measured platelet deposition on PTFE grafts after balloon catheter thrombectomy as an indicator o f the likelihood of rethrombosis. MATERIAL AND METHODS Three different canine models of PTFE graft thrombosis were created: (1) in vitro: PTFE grafts filled with autologous blood, clamped at either end, allowed to clot, and stored in sterile normal saline solution at 37 ° C; (2) in vivo: PTFE grafts filled with
Volume11 Number6 June1990
Platelet deposition on PTFE grafts after thrombectomy
I 1, 2 and 3 Week
Fig. 1. Three different models of graft thrombosis were created weekly in each dog for 3 weeks, resuking in nine thrombosed grafts per dog. In sire grafts were anastomosed end to end to the femoral or carotid artery and oversewn distally. In vivo grafts were filledwith blood, oversewn at both ends, and placed in the subcutaneous tissue near the in situ grafts. In vitro grafts were filled with blood, clamped, and placed in a saline water bath. :ut :logous blood, oversewn at each end with polypropylene suture, and implanted in a subcutaneous pocket; and (3) in situ: one end of PTFE grafts anastomosed end to end to a femoral or carotid artery with 6-0 polypropylene suture and the other end oversewn as a blind tube. Duration of thrombosis was varied from 1 to 3 weeks by initiating one graft of each model type in each dog (eight female dogs from random sources) at weekly intervals for 3 weeks and then harvesting the grafts simultaneously. Graft implantation (in vivo and in situ types) was performed by use of sterile surgical technique without anficoagulation under anesthesia induced with thiamylal sodium and maintained with halothane and nitrous oxide inhalation. Identical lengths of expanded PTFE grafts were used in a~ models (expanded PTFE, 4 mm inside diamter by 7 cm in length, internodal distance 30 I~m, 0.39 mm thick; W. L. Gore & Associates, Inc., Elkton, Md.). To accommodate the three in situ grafts, both femoral arteries and one carotid artery served as sites of in situ graft placement during successive weeks. In vivo grafts were implanted in subcutaneous pockets associated with these incision sites (Fig. 1). After 3 weeks, animals were reanesthetized with intravenous pentobarbital and mechanically ventilated. The six thrombosed grafts from each dog were harvested (three in vivo and three in situ), which together with the three in vitro grafts represented thrombosis of i, 2, and 3 weeks' duration for each of the three models. Graft thrombectomy was then performed by making three extractions with a 3F
balloon catheter (Fogerty, model 12-080-3F; American Edwards Laboratories, Anasco, Puerto Rico) inflated to a uniform size that simulated clinical thrombectomy techniques. Indium 111-labeled ptatelet deposition was then measured in each graft using an ex vivo flow system in the dog from which the grafts were harvested. A tandem flow circuit was created as described previously5 by placing 4 mm outside diameter by 3 cm long polyethylene cannulas into each femoral artery. The PTFE grafts were attached to these and similar distal cannulas with simple ligatures, exposing the middle 5 cm of graft to flowing blood. Effluent blood from each of the grafts were routed in silicone rubber tubing through a tandem roller pump and an in-line electromagnetic f l o w probe to maintain a constant flow rate of 55 ml/min through each graft. Blood was returned to the animal through a femoral venous cannula. Heparin (300 units/kg intravenously) was given before the ex vivo circulation was established. Autologous platelets were labeled with indium 111 oxine as described previously6 and readministered to each dog before ex vivo graft perfusion. In each dog the nine thrombectomized grafts, as well as three control grafts (new PTFE grafts, not previously exposed to blood), were perfused with platelet-labeled blood for 20 minutes as a series of six graft pairs in the tandem flow circuit. The order of graft perfusion was deliberately varied between dogs to prevent sequence effects. We have previously demonstrated that repetitive perfusion of control grafts in this model results in uniform platelet deposition during the time course used in this study, s
Journal o f VASCULAR
806 Bech,Galt, and Cronenwett
b N 2.o
Fig. 2. Platelet deposition on grafts briefly exposed to plasma and blood (n = 3 dogs) and on thrombectomized grafts of each thrombosis model type (n = 8 dogs) expressed as the ratio ( -+SEM) of the study grafts to the control (new) grafts. Hatelet deposition was measured by "1In-labeled platelet counting normalized for circulating blood. Thrombectomized graft data represent the mean of 1-, 2-, and 3-week thrombus duration. All three thrombectomized graft models accumulated significantly more platelets than did new control grafts (p < 0.05).
Table I. Platelet deposition on thrombectomized and control PTFE grafts Duration ofgraft thrombosis (rain cpm per grafi/~llIn cpm per 100 Izl of circulating blood) l~iodel
Sequence of controlgraJ~perfusion (~In cpm per graft/mIn cpm per 100 td of circulating blood)
In situ In vitro
5.8 + 2.0 6.2 + 2.4
6.2 + 3.3 4.2 + 1.4
7 . 5 -+ 2 . 6 8 . 3 -+ 2 . 6
In vivo Control*
8.4 + 2.9 -
13.4 + 5.3 -
7.5 ± 1.8 -
3 . 7 -+ 1 . 6
2.7 + 0.7
2.4 + 0.6
Data are mean + SEM; n = 8 dogs each group. *Identical new grafts not previouslyexposed to blood.
Control grafts were assigned to pairs at the beginning, middle, and end of the sequence, and the nine thrombectomized grafts were randomly distributed among them. Blood from the circuit was sampled twice during each perfusion (2 ml per sample) as an index of circulating platelet activity. At the conclusion of each perfusion, grafts were rinsed for 2 minutes with normal saline solution using a separate roller pump at the same flow rate. Each graft was then cut into 1 cm segments. The middle five segments were placed in scintillation vials and counted with blood samples in a -/-well counter for 5 minutes each (the ends not exposed to blood were discarded). Platelet deposition on each segment was calculated as the 111In counts/min/cm graft/Hlln counts/min/100 Ixl of circnlating blood. Total graft platelet deposition was calculated as the sum of the five graft segments.
To examine the effect of brief exposure to autologous blood or plasma on graft-platelet deposition, three additional dogs were studied with identical ex vivo flow circuits. In each dog four PTFE grafts were filled with blood and four grafts were filled with plasma. The grafts were clamped for 1 hour in a 37 ° C water bath and then evacuated as above with a balloon catheter (as if performing a thrombectomy) before perfusion in the tandem flow circuit with four new (control) PTFE grafts. Before inclusion in this study, all dogs were prescreened for platelet aggregation response to 0.54 ~ m o l / L epinephrine-enhanced arachidonic acid (1 mmol/L) as described previously. 6 Only animals with a strongly positive aggregation response (>50% change in light transmission) were included in the study. Animals were cared for in accordance with the "Principles ofLaboratory Animal Care" (formulated by
Volume 11 Number 6 June 1990
Platelet deposition on PTFE grafts after thrombectomy 807
4.0 E O o
1.0 1 Week
Duration of Thrombosis
Fig. 3. Platelet deposition on thrombectomized grafts (n = 8 dogs) according to the duration of thrombosis (mean of three thrombosis models) expressed as the ratio (-+ SEM) of thrombectomized to control grafts. The tendency for increased platelet deposition from 1 to 3 weeks was not statistically significant.
the National Society for Medical Research) and "Guide for the Care and Use of Laboratory Animals" (?~ational Institutes of Health publication number 80-23, revised 1985). Data were analyzed for statistical significance by analysis of variance. RESULTS
At explanation the three different models of graft thrombosis bore characteristic gross appearances. In vitro grafts removed from the saline bath remained white, supple, and 7 cm in length. Thin liquid thrombus was present within the lumen. In vivo grafts removed from the subcutaneous tissues were yellowgray, stiff, and contracted, measuring 6 cm long. Dark gelatinous thrombus was present in some cases, with more serous material in others. In situ grafts were also stiff and contracted to 6 cm. Dense white thrombus was firmly adherent at the oversewn blind ~.,d, and both red and white thrombus were present throughout the remainder of the graft. This thrombus was the most difficult to remove completely with a balloon catheter regardless of the duration of thrombus (1, 2, or 3 weeks). Platelet deposition was greater on all thrombectomized grafts compared with new (control) grafts at each time point (1, 2, and 3 weeks; Table I). Expressed as ratios of thrombectomized to control grafts, each model type demonstrated significantly increased platelet accumulation independent of thrombus duration: in vivo, 3.7 + 0.6 times control (+SEM; p < 0.01); in situ, 2.6 + 0.4 times control (p < 0.05); and in vitro, 2.0 _+ 0.3 times control (p < 0.05). In contrast, 1 hour of exposure to either blood or plasma resulted in statistically insignificant increases in platelet deposition (both 1.3 times control; Fig. 2). Although a trend toward increasing platelet deposition with increased duration
of thrombosis was seen, this was not statistically significant (Fig. 3). Deposition ofplatelets on the 1 cm segments of in vitro, in vivo, and control grafts was relatively uniform from the proximal to distal flow ends. More platelets were deposited near the blind end (distal segment in the perfusion circuit) of in situ grafts than on other segments (Fig. 4). DISCUSSION
Balloon catheter thrombectomy has successfully improved the secondary patency rate ofinfrainguinal PTFE grafts but may also result in early graft rethrombosis. 7,8This occurs not only because of limited outflow but also when no underlying correctable causc is found for the initial thrombosis. Strept et al.9 found no anatomic explanation for 48% of failed primary infrainguinal grafts and recommended subsequent anticoagulation for these patients. Although considerable evidence exists that early primary graft occlusion is mediated by platelet deposition, 3'4'1° this effect has not been studied after graft thrombectomy. Despite the logical assumption that newly thrombectomized grafts are more thrombogenic than are fresh grafts, we could not find published evidence for this nor could we find an animal model of graft thrombosis that has been developed to study this problem. We investigated three methods of artificially creating PTFE graft thrombosis with the intention of developing an animal resource-sparing model to study graft thrombectomy. All three models of thrombosis resulted in a twofold to threefold increase in immediate platelet deposition after thrombectomy, compared with new grafts. Because this occurred in a perfusion system that controlled blood flow rate and platelet reactivity, the data confirm our clinical impression that a newly thrombectomized
Journal of VASCULAR SURGERY
Bech, Galt, and Cronenwett
In vitro Grafts
In vivo Grafts
In situ Grafts
Fig. 4. Platelet deposition on 1 cm segments of control and thrombectomized grafts (n = 8 dogs; mean of 1-, 2-, and 3-week thrombus duration). Segments A to E represent proximal to distal segments in the flow circuit for all grafts. Segment A in in sire grafts represents the anastomotic end of the graft animal, whereas segment E represents the blind end. Platelet deposition is expressed as rain cpm per cm graft/1HIn cpm per 100 ~1 of circulating blood. PTFE graft surface is particularly attractive to circulating platelets. The graft surface encountered by blood after reestablishing flow depends on a number of factors. The age of the graft will have influenced the degree and character of primary "healing" that occurred before thrombosis, imparting a thrombotic potential that contributed to the primary occlusion. The restored flow surface after thrombectomy is unlikely to be less thrombogenic. Both thrombolytic and inflammatory processes alter the character of the thrombus within the graft over time, resulting in thrombus organization that may prevent complete thrombectomy. Furthermore, a balloon catheter can induce spasm or intimal damage in the artery near an anastomosis, cause distal embolization, or incompletely remove thrombus, leaving an irregular graft surface that may foster locally turbulent blood flow and more platelet deposition) ~ Physical characteristics of newly implanted PTFE grafts theoretically reduce thrombogenicity. These include a smooth, hydrophobic luminal surface, chemical inertness, and electronegativity)2 Despite these attributes, proteins adhere rapidly to the PTFE surface and attract platelet aggregates within 60 seconds of exposure to blood) 2 The adsorption of pro-
tein, particularly fibrinogen, to artificial graft surfaces has by itself been shown to enhance the adherence of platelets and the initiation of thrombus formation. 13 Kenny et al.14 demonstrated that 64% of the surface of PTFE grafts implanted in dogs was covered by thrombus after only 6 hours. The interstices of canine expanded PTFE grafts are filled with fibrin, red and white blood cells, platelets, and macrophages by 2 to 4 weeks after implantation. ~ Although if, timal pannus ingrowth occurs at the anastomoses, a thin acellular fbrin coagulum exists along the majority of the flow surface for an extended period in dogs and indefinitely in humans. It is on this surface that thrombus forms, and it is this surface, at best, that is restored after thrombectomy. Because thrombectomy is unlikely to be perfectly complete, an irregular overlay of retained thrombus would confer additional thrombogenicity as a result of both the thrombus itself and increased surface roughness.l°,16 Although our three models of graft thrombosis attracted more platelets than did control grafts, they also showed individual differences. In vivo grafts attracted the most platelets, in a homogeneous distribution. Ligation of both graft ends theoretically prevented migration of inflammatory cells, contributing
Volume 11 Number 6 June 1990
Platelet deposition on PTFE grafts after thrombectomy 809
to the lack of extensive thrombus organization that was observed. Platelet deposition after reperfusion of these grafts is most easily attributed to the presence of a protein coagulum within the graft wall and on its flow surface, the latter acellular and relatively unchanged by the balloon catheter. In vitro grafts attracted the fewest platelets of the three models of thrombosis. When thrombectomized, they contained a thin serosanguinous fluid, suggesting that the thrombus had been diluted by saline solution permeating the graft wall during storage. In contrast to the other models, these grafts maintained their native shape, suggesting that they were not thoroughly suffused with a protein coagulum. Nonetheless, they did accumulate significantly more platelets than did control grafts. Because brief (1 hour) contact with blood or plasma did not substantially increase platelet deposition, it appears that more prolonged exposure to plasma protein is required for a significant surface ~- zration to occur. In situ grafts attracted more platelets at the oversewn end than elsewhere, where white (platelet) thrombus was densely adherent before thrombectomy. Retained thrombus in this region probably contributed to increased platelet deposition in the distal segments of these grafts. Less accumulation of platelets overall by in situ compared with in vivo grafts could reflect a variably delayed thrombosis of the former type as a result of circulating blood in the proximal graft, thus decreasing the actual duration of the thrombus. The ideal animal model to study PTFE graft thrombectomy would use a graft exposed to normal circulation that thrombosed spontaneously. However, such a model is severely limited by the lack of uniformity in the timing of graft thrombosis after ufiplantation. We abandoned this approach because it was impossible to standardize the duration of thrombosis and previous circulation exposure. We therefore tested three models in which the onset of thrombosis was known precisely. Our in vitro model was attractive because it does not require animal surgery and any number of grafts can be tested. However, the gross appearance of these grafts and thrombus bore little resemblance to a clinically thrombosed graft, and the least platelet deposition occurred. Conversely, the in situ model resulted in graft thrombus that was most similar to clinical experience, with dense, adherent thrombus and increased platelet deposition in that region. However, this model restricts the number of grafts that initially can be thrombosed per animal and subsequently reimplanted in vivo because of the limited number of arterial sites in each dog. It is also more time and resource intensive, be-
cause it requires the initial anastomosis that may introduce some thrombosis variability. The in vivo model was selected as a compromise that exposed grafts to the subcutaneous milieu but avoided arterial disruption so that subsequent in vivo implantation ofthrombectomized grafts could be performed in the same animal. This technique resulted in the highest, most uniform platelet deposition in grafts that took on the appearance of clinical thrombosis. We believe that our in vivo model is potentially the most useful for subsequent experiments. There is some controversy regarding the effect of thrombus age on the outcome of graft thrombolysis.17'ls In our study we found that thrombogenicity reflected by platelet deposition did not increase significantly from 1 to 3 weeks. Curl et al.~8 found that the infusion time required to lyse graft thrombus with three different thrombolytic agents increased with the age of thrombus during 1 to 7 days but found no difference in thrombus-free surface area after thrombolysis, suggesting that completeness of thrombus removal did not vary with thrombus duration. We are not aware of data that have described the effect ofthrombus duration on successful surgical thrombectomy. Our results suggest that 1-week implantation is sufficiently long to create a useful model to study increased thrombogenicity after thrombectomy. Although our study demonstrated increased ex vivo platelet deposition immediately after graft thrombectomy, we can only speculate concerning the clinical significance of this observation and its association with increased rethrombosis. This hypothesis can be tested in vivo by implanting thrombectomized grafts in dogs and measuring subsequent patency. Our initial data suggest, however, that thrombectomized grafts should be considered to be at high risk for early reocclusion regardless of structural or flow-related constraints on their performance. We emphasize that increased platelet deposition was demonstrated in the first 20 minutes after reexposure to blood, a finding we believe supports the need for antiplatelet therapy beginning at or before the time of thrombectomy, as others have advocated39 Unfortunately, this strategy is currently hampered by the lack of clinically available intravenous antiplatelet agents that can be given during surgery with immediate effect. Our canine model may be useful for testing the effect of such drugs on early graft rethrombosis after thrombectomy. Heidi Clough provided expert technical assistance. I~FE grafts were provided by W.L. Gore & Associates, Inc., Elkton, Md.
lournal of VASCULAR SURGERY
810 Bech, Gall, and Cronenwett
REFERENCES 1. Ascer E, Collier P, Gupta SK, Veith FJ. Reoperation for polytettafluoroethylene bypass failure: the importance of distal outflow site and operative technique in determining outcome. J VAse Suv~G1987;5:298-310. 2. Veith FJ, Gupta S, Daly V. Management of early and late thrombosis of expanded polytetrafluoroerhylene (PTFE) femoropopliteal bypass grafts: favorable prognosis with appropriate reoperation. Surgery 1980;87:581-7. 3. Callow AD, Counolly R, O'Donnell TF Jr, et al. Plateletarterial synthetic graft interaction and its modification. Arch Surg 1982;117:1447-55. 4. Mackey WC, Keough EM, Connolly RJ, et al. A baboon flow-regulated shunt for the study of small-caliber vascular grafts. J Surg Res 1984;37:112-8. 5. Boorstein IM, Endean ED, Cronenwett JL. Intra-artetial carbacyclin infusion inhibits canine platelet deposition on PTFE grafts. Trans Am Soc ArxJfIntern Organs 1986;32:360-2. 6. McDaniel MD, Huntsman WT, Miett TOC, Cronenwett JL. Effect of a selective thromboxane synthase inhibitor on arterial graft patency and platelet deposition in dogs, Arch Surg 1987;122:887-92. 7. BakerWH, Hadcock MM, Littooy FN. Management of polytetrafluoroethylene graft occlusions. Arch Surg 1980;115: 508-13. 8. Green RM, Ouriel K, Ricotta JJ, DeWeese JA. Revision of failed infrainguinal bypass graft: principles of management. Surgery 1986;100:646-53. 9. Stcpt LL, Flinn WR, McCarthy WI III, Bartlett ST, Bergan JJ, Yao JST. Technical defects as cause of early graft failure after femorodistal bypass. Arch Surg 1987;122:599-604. 10. Sicard GA, Allen BT, Lee K, et al. Platelet reactivity as a fimction of thrombus-frce surface in small-diameter vascular grafts. ASAIO Trans 1985;8:86-9. 11. Barone GW, Conerly JM, Farley PC, Flanagan TL, Kron IL.
Endothelial injury and vascular dysfimction associated with the Fogarty balloon catheter. J VASCSURG1989;9:422-5. Boyce P. Physical characteristics of expanded polytetrafluoroethylene grafts. In: StanleyJC, Burkel WE, Lindenaner SM, Bartlett RH, Turcotte JG, eds. Biologic and synthetic vascular prostheses. New York: Grune & Stratton, 1982:553-61. Barber TA, Mathis T, Ihlenfeld JV, Cooper SL, Mosher DF. Short-term interactions of blood with polymeric vasculargraft materials: protein adsorption, thrombus formation, and leukocyte deposition. Scanning Electron Microsc 1978;2:43140. Kenny DA, Berger K, Walker MW, et al. Experimental comparison of the thrombogenicity of fibrin and PTFE flow surfaces. Ann Surg 1980;191:355-61. Graham LM, Burkel WE, Ford JW, Vinter DW, Kahn RH, Stanley JC. Expanded polytetrafluoroethylene vascular prostheses seeded with enzymaticallyderived and cultured canine endothelial cells. Surgery 1982;91:550-9. Zingg W, Neumann AW, Strong AB, Hum OS, Absolom DR. Effect of surface roughness on platelet adhesion under static and under flow conditions. Can J Surg 1982;25:16-9. HaUett IW Jr, Greenwood LH, Yrizarry JM, Pierson WP, Robison JG, Brown SB. Statistical determinants of success and complications of thrombolytic therapy for arterial ocd_"~ sion of lower extremity. Surg Gynecol Obstet 1985;161: 431-7. Cud GR, Jakubowski JA, Nabseth DC, Bush HL Jr. Efficacy of tissue plasminogen activator and urokinase in a canine model of prosthetic graft thrombosis. Arch Surg 1986; 121:782-8. Collier P, Ascer E, Veith FJ, Gupta SK, Nunez AK. Acute thrombosis of arterial grafts. In: Bergan JJ, Yao JST, eds. Vascular surgical emergencies. 1st ed. Orlando: Grune & Stratton, 1987:517-28.
Dr. Thomas F. O'Donnell (Boston, Mass.). You have addressed a clinically relevant problem, because we frequently are confronted with thrombosis of polytetrafluoroethylene (PTFE) grafts either in the above-knee position or as an axillofemoral bypass. Our experience, like that of the Montefiore group, has shown that rates of above-knee PTFE grafts can be improved 15% to 20% by graft thrombectomy and correction of the cause of graft failure. What intrigues me about this study is the model, and I would like to focus on that. You have used indium-labeled platelets as the means of assessing the thrombogenieity of the graft material. Bill Mackey and Jens-Jorgansen have studied the effect of both flow rates and type of graft material on indium-labeled platelet uptake and have shown that it certainly relates to the type of graft material. PTFE is less reactive with platelets than is Dacron. In addition, indiumlabeled platelet uptake is related to the length of time after implantation, particularly when that graft is in a circuit connected with flowing blood. Finally, platelet uptake var-
ies with the portion of the graft sampled, so you were ver~"~ wise to avoid in the sampling technique the ends of the grafts that are usually the most thrombogenic and platelet reactive. When one addresses a clinical problem in the laboratory, however, the experimental model should closely parallel what one confronts clinically. That flaw in the study design is what bothers me a little bit about this study. The model does not permit the usual development of pseudointima that will occur in humans. Certainly in dogs one must be cognizant of complete endothelialization, which in a 7 cm PTFE graft may occur within 6 weeks. A second problem with the model is the lack of exposure of the graft to a flow circuit that can produce thrombus. Would it not be better to produce thrombus in a circuit than to perform thrombectomy? Second, what did histologic examinations of your grafts show? Did the three types that you used-in vitro, in vivo, and in situ--vary, because certainly that will determine the amount of platelet deposition? Finally, many of us now employ graft thrombectomy with balloon
Volume 11 Number 6 June 1990
Platelet deposition on PTFE grafts a~er thrombectomy 811
catheters for failed PTFE grafts selectively, prefering lytic agents for chemical thrombectomy. With your interest in lyric therapy, could you hypothesize or perhaps provide some data on what the difference would be if you lysed a graft chemically with urokinase and then measured platelet uptake. At New England Medical Center we have seen that more than 90% of failed above-knee PTFE grafts can be lysed with urokinase. Dr. F. R. Bech (closing). We were interested in developing an animal model of this problem chiefly because we could not find another one published in the literature. We started our investigations by implanting grafts in dogs and awaiting spontaneous thrombosis but found that it was impossible to predict exactly when the grafts thrombosed. Because we were interested in the question of whether duration of thrombosis affected platelet deposition, we needed to know exactly when the grafts did throm-
bose. So after several attempts at a model of spontaneous thrombosis, we abandoned it in favor of the current model. There are little data to suggest that in the first few weeks after implantation of PTFE grafts in a dog, that a real pseudointimal layer is developed. At 6 weeks a sparse population of cells accumulates, but in the acute period after implantation a new graft is coated with only a protein coagulum, specifically fibrin. We have not yet studied the histologic factors of these grafts, and the photomicrographs shown today represent only an initial view of the thrombectomized grafts surface. We intend to perform histologic analysis and characterize biochemically the protein coagulum left behind after thrombectomy. Finally, lyric therapy is an interest of ours and we intend to use this model to try to answer the question of whether the protein coagulum might be eliminated or reduced by different lytic agents.