- Email: [email protected]
Intraoperative fibrinolytic therapy: Experimental evaluation
Intraoperative fibrinolytic therapy: Experimental evaluation William J. Quifiones-Baldrich, M.D., Stanley Ziomek, M . D . , T h e o d o r e C. H e n d e r s o n , B.A., and Wesley S. Moore, M.D., Los Angeles,
Calif.
Percutaneous intra-arterial infusion of fibrinolytic agents has emerged as an alternative to embolectomy in selected patients with acute arterial occlusions. The combination of fibrinolytic therapy and embolectomy may be superior to either modality alone. This experiment was designed to determine safety and efficacy of intraoperative fibrinolytic therapy as an adjunct to catheter embolectomy. Forty hind limbs in 20 adult mongrel dogs were embolized with thrombus created in vitro. After 24 hours, bilateral transfemoral embolectomy was followed by an intra-arterial, intraoperative infusion. Fifteen limbs (control) received 250 ml of saline solution during a 30-minute Period; 25 limbs (experimental) received an arterial infusion of 60,000 units of streptokinase during a 30minute Period (SK 30'). In five limbs of each group, 500 units of heparin (H) was added. In five experimental limbs the streptokinase infusion time was increased to 60 minutes (SK 60'). Arteriograms and blood flow measurements were obtained before and after embolectomy (PE) and after infusion (PI); the results were compared. Improvement between the PE and PI angiograms was seen in 20% (3 of 15 dogs) of control subjects. In contrast, improvement after the infusion was evident in 100% (5 of 5 dogs) of dogs given SK plus H 30' (p < 0.01), in 80% (12 of I5 dogs) of dogs given SK 30' (p < 0.01), and in 20% (1 of 5 dogs) of dogs given SK 60'. A trend toward increased blood flow was noted in the experimental group. There were no intraoperative complications with hemostasis or postoperative bleeding (36-hour observation). We conclude that intraoperative fibrinolytic therapy in dogs is safe and effective as an adjunct to thromboembolectomy. A human clinical trial is recommended. (J VASC SURG 1986; 4:229-36.)
Catheter thromboembolectomy remains the initial procedure of choice in patients with thromboemboric complications of atherosclerosis. The intra-arterial infusion of fibrinolytic agents by percutaneous catheterization has emerged as an alternative to catheter thromboembolectomy in a selected group of patients. The results obtained with fibrinolytic therapy have been variable with partial lysis achieved in most instances and most patients requiring corrective surgery. ~-4 Experimental studies of hind limb ischemia have suggested the occurrence of arteriolar thrombosis in muscular and skin vessels after 6 hours of complete inflow occlusion/ Delays in surgical embolectomy may be hampered by similar peripheral clot propaFrom the Department of Surgery, Universityof California, Los Angeles, Center for the Health Sciences,and the ResearchDivision of the SepulvedaVeterans AdministrationMedicalCenter, Sepulveda,Calif. Supported in part by the Career DevelopmentAward Program, Veterans Administration. Reprint requests: WilliamJ. Quifiones-Baldrich,M.D., Assistant Professor of Surgery, Department of Surgery, Sectionof Vascular Surgery,UCLA Center for the Health Sciences,Los Angeles, CA 90024.
gation, which may be inaccessible to the embolectomy catheter. Greep et al.6 was able to recover additional thrombotic material in 110 peripheral arterial embolectomies with the use of a Dormia catheter in "almost all" cases after thorough Fogarty catheter embolectomy. The only clinical study that has addressed this issue with routine angiography after embolectomy reported a 36% incidence of incomplete embolectomy. 7 Therefore, incomplete removal of all intravascular clot during embolectomy is a relatively frequent problem. Bleeding complications of intra-arterial fibrinolytic therapy are a result of systemic effects of the agent, which in turn are related to the prolonged infusion time required to achieve complete resolution of the process. With these facts in mind, we postulate a role of fibrinolytic agents as an adjunct to catheter thromboembolectomy. The idea of removing the bulk of thrombus surgically and of lysing any remaining defect is attractive from the therapeutic standpoint. The purpose of these experiments is to investigate the safety and effectiveness of fibrinolytic therapy as an adjunct to catheter thromboembolectomy. 229
230 Qui~ones-Baldrich et al.
MATERIAL A N D M E T H O D S
The model of thromboembolism described in these experiments uses true thrombus. The distinction between thrombus and clot is particularly important in a canine model. Embolization of fresh clot in the canine circulation results in temporary arterial occlusion, followed by rapid lysis and dissolution because of a very active canine fibrinolytic system. Clot consists of all elements of blood, mainly red cells bound by a fibrin mesh. In contrast, thrombus consists mainly of platelets in a fbrin matrix with few red and white cells. Clot is formed by allowing coagulation to proceed in stagnant blood. Thrombus is produced when coagulation proceeds while blood is in linear motion. This concept was originally proposed by Chandler8 who described a method to produce thrombus in vitro. An aliquot of fresh blood is introduced in a 25 cm length of polyvinyl tubing and rotated in a 23-degree slanted turntable at 16 rpm. Coagulation proceeds while blood is in motion at a linear velocity of 400 cm/min. After 20 minutes, a small piece of thrombus forms at the forward tip, with the remaining blood consisting of a tail of clot or liquid from defibrinization. Using this method we have produced in vitro thrombus and embolized this material in the hind limbs of dogs. Preliminary studies in our laboratory demonstrated that embolization with thrombus produces a stable, permanent occlusion in the canine hind limb. This was compared in the same animals with clot embolization. Repeat angiography showed continued resolution with eventual complete clearing in every instance in those limbs embolized with clot. Twenty adult mongrel dogs (weighing 15 to 20 kg) of either sex underwent embolization with thrombus through a No. 7 Fr. single-lumen angiographic catheter introduced through a left carotid vessel cut down and placed under fluoroscopic guidance into the superficial femoral artery. All procedures were done with the dogs anesthetized with intravenous thiamylal (Surital), 11 mg/kg. This research was conducted under guidelines of the NIH publication "Guide for Care and Use of Laboratory Animals." Both right and left superficial femoral arteries were embolized with 4.5 to 5 ml of thrombus each. Angiography was done through the same catheter once retracted to the infrarenal aorta. An automatic film changer was programmed for two films per second for 2 seconds for the pelvis and thigh views and two films per second for 4 seconds for the distal runoff with a 3- and 4-second injection delay, respectively. Injection was achieved with a pressure
Journal of VASCULAR SURGERY
injector at 600 psi with 20 ml of 76% meglumine for the pelvis and thigh views and 30 ml of 76% meglumine for the infrageniculate runoffviews. Aortography with runoff was performed before and immediately after embolization. Twenty-four hours after embolization, transfemoral embolectomy was accomplished in all 40 limbs with the dogs placed under general endotracheal anesthesia with halothane (Fluothane) and with mechanical ventilatory support. With the use of sterile technique, both femoral arteries were exposed. Femoral blood flow was measured with an electromagnetic flowmeter (Gould Statham SP2204) before embolectomy. Three separate readings with an appropriately sized probe placed around the common femoral artery were made and recorded. Through a transverse arteriotomy in the common femoral artery, No. 3 and No. 4 Fogarty catheters were introduced distally 35 to 45 cm depending on the limb length. Embolectomy maneuvers were continued until two consecutive passes failed to retrieve any material. The amount of thrombus removed was recorded. The arterotomies were closed with interrupted 6-0 polypropylene sutures, after back-bleeding the vessels. Flow was then reestablished. Heparin (100 U/kg) was given intravenously before embolectomy. Topical bovine thrombin was used for hemostasis when necessary. Femoral blood flow measurements were taken immediately after embolectomy and recorded. Angiography was repeated immediately after completion of embolectomy. Following angiography after embolectomy, an intra-arterial infusion of either control or experimental solution was administered. A 22-gauge Teflon angiocatheter was inserted in the artery distal to an occluding vascular clamp (Fig. 1). The relationship of the arteriotomy to the occluding clamp was variable; however, the arteriotomy was always proximal to the catheter. Control solution consisted of 250 ml of normal saline solution in 10 limbs and 250 ml of normal saline solution with 500 units of heparin in five limbs. Experimental solution consisted of 250 ml of normal saline solution with 60,000 units ofstreptokinase (Streptase MSD) in 15 limbs and 250 ml of normal saline solution with 60,000 units of streptokinase and 500 units of heparin in five extremities. The infusion was given during a 30-minute period in 35 limbs (15 control and 20 experimental) and a 60-minute period in five experimental limbs (Fig. 2). Control and experimental limbs were chosen at random in the first five animals. Experimental and control limbs were chosen in the remaining 15 dogs according to results after cmbolectomy, assigning to
Volume 4 Number 3 September 1986
Intraoperative fibrinolysis: Experimental evaluation 231
/
I0- Saline
15 CONTROL
~
5- Saline + Heparin
4 0 HIND LIMBS
/
15-60,000 U SK; 50'
25 EXPERIMENTALt5 -60,000 U SK+ Heparin; 50' k - - 5 - 6 0 , 0 0 0 U SK; 60'
Fig. 2. Control and experimental groups. In 15 animals (30 limbs) one extremity was assigned to the control group and the opposite to the experimental group. In five animals (10 limbs) both extremities were assigned to the experimental group. In the latter, the right limb was infused during a 60-minute period, and the left during a 30-minute period. ~duoJ ~abus
Fig. 1. Drawing of normal arterial anatomy of canine hind limb. Note infusion site is distal to arteriotomy and an occlusive vascular clamp. the experimental group those limbs with angiographic evidence of intraluminal defects. In those animals in which both limbs had residual occlusions, the worst o f the two limbs was assigned to the experimental group. In five animals, both limbs were assigned to the experimental group, giving the intraoperative infusion after embolectomy during a 30minute period in the right limb and a 60-minute period in the left limb. Femoral blood flow measurements were done immediately after infusion and the average of three readings recorded. The angiographic catheter in the aorta was used to measure central blood pressure during blood flow measurements. An attempt to maintain the animal's blood pressure within 15 mm H g of the baseline was made. However, there were instances in which this could not be achieved and the pressure was higher or lower than baseline. Certainly this introduces an error in the absolute value o f any one particular blood flow measurement; however, in all
animals except five, one limb was control and one experimental so that the relative difference between the two sides retains validity. Angiograms were reviewed in a blinded fashion and considered without knowledge of the type of infusion administered. Specifically, changes between the angiograms performed before and after embolectomy (PE) and after infusion (PI) were noted. Improvement was considered as the reappearance of vessels not seen after embolectomy and/or disappearance of intraluminal defects causing partial restriction of flow as seen on the angiogram done after embolectomy. If no change was seen between these two studies, the result was recorded as unchanged. If the PI angiogram failed to show vessels that were seen on the angiogram after embolectomy, the result was recorded as worsened. When possible, angiography was repeated between 24 and 48 hours after the infusion. However, the total contrast load given to these animals precluded survival for this length of time. In future studies, we would modify the protocol and eliminate both the baseline and the angiogram after embolization, because the former was useful in only one instance and the latter did not add any valuable information. RESULTS The baseline angiogram revealed normal anatomy in 38 of 40 limbs. Two limbs in the same animal showed the popliteal artery ending in small branches below the knee and the distal circulation supplied by the saphenous artery. Embolization with thrombus achieved an occlusion of the superficial femoral artery in all instances. Transfemoral embolectomy retrieved thrombus and clot in all instances, the average recovered being 3 + 1 ml. The Fogarty balloon catheter could be introduced for the length of the extremity in 38 limbs. In two limbs, an anomalous popliteal artery as seen in the baseline angiogram
Journal of VASCULAR SURGERY
232 Qui~ones-Baldrich et M.
Fig. 3. A, Angiogram after bilateral transfemoral embolectomies shows saphenous artery and distal tibial artery occlusion in right limb (curped and lade stra~qht arrow). Left system appears intact except for some defects in thigh branches. B, Angiogram after experimental 30-minute infusion (no heparin) in the right limb and control infusion in the left limb. Note significant improvement in both saphenous and tibial arteries on right. The left (control) distal tibial artery is now occluded with reconstitution of the pedal arch by collateral circulation. precluded passage of the catheter beyond the infrageniculate region. Residual clot and thrombus after embolectomy was seen in 34 of 40 (85%) limbs on angiograms done after embolectomy. Embolectomy was complete in six limbs (four control and two experimental). Improvement between the PE and PI angiograms was seen in 2 of i0 control limbs infused with 250 ml of saline solution. In this group, seven limbs were unchanged and one was worse after infusion (Fig. 3). One of five control limbs infused with saline solution and heparin appeared improved, whereas one was worse and three unchanged after infusion. All control limbs were infused over 30 minutes. Thus, a total of 3 of 15 control limbs showed improvement between the PE and PI angiograms (Fig. 4) with no difference seen between the group that received saline solution alone and that which received saline solution
plus heparin. In contrast, improvement between the PE and PI angiograms was seen in 12 of 15 limbs infused over 30 minutes with 60,000 units of streptokinase dissolved in 250 ml of saline solution. In two limbs there was no appreciable difference between the PE and PI angiograms. In one limb the PI angiogram was worse, showing reocclusion of the superficial femoral artery. The difference between the control (3 of 15 dogs) and experimental (12 of 15 dogs) groups is statistically significant (p < 0.003) by Fisher's exact two-tailed test. Five experimental limbs were infused with a solution of 60,000 units ofstreptokinase and 500 units of heparin in 250 ml of normal saline solution. PI angiograms revealed improvement in all five extremities (Fig. 4). This is a statistically significant difference when compared with the control group (p < 0.004); however, it is not significant when
Volume 4 Number 3 September 1986
Intraoperative fibrinolysis: Experimental evaluation 233
Fig. 4. A, Angiogram done after embolectomy shows residual clot bilaterally (arrows). B, After control infusion with heparin in right limb residual clot is unchanged, with some improvement in the middle digital artery. After 60,000 units of streptokinase with 1000 units of heparin was infused there is remarkable improvement in all three digital arteries with possible migration of intravascular clot (left arrow). Both results were recorded as improved. compared with the experimental group receiving no intra-arterial heparin in conjunction with the streptokinase infusion. In five experimental limbs, the infusion time was 60 minutes, the infusate consisting of 60,000 units of streptokinase in 250 ml of normal saline solution. Improvement was seen in one of the five limbs. In the remaining four limbs no appreciable differences between the PE and the PI angiograms were seen. In these animals, the opposite limb was infused over 30 minutes with 60,000 units of streptokinase in 250 ml of normal saline solution. Angiography showed improvement in three limbs, one was unchanged, and in one worsening of the PI appearance was evident. The 30-minute results are included in the first experimental group already presented. The results of blood flow measurements before and after embolectomy and after the intraoperative intra-arterial infusion are shown in Fig. 5. The mean + 2 standard deviations of the blood flow after embolectomy was 41.1 _ 6.4 ml/min for control and 30.7 _+ 5.0 ml/min for experimental limbs. This lower figure in the experimental limbs is the result of selecting the worst of the embolized limbs for the experimental group. The mean blood flow after infusion was 49.9 _+ 6.4 ml/min for control and 47.4 + 6.3 ml/min for experimental limbs. The mean increase (A) of 8.3 ml/min in the control group
vs. 16.7 ml/min in the experimental group does not achieve statistical significance (p < 0.13). However, this is likely the result of the small number of experiments and suggests a positive trend toward increased blood flow in the experimental group. When possible, angiography was repeated 24 to 48 hours after infusion. Twelve animals died within the 24 hours after embolectomy and infusion. Complete autopsy was performed in the first five deaths; no abnormalities were found to explain the outcome. Specifically, no intra-abdominal or intrathoracic bleeding was present. Most of these animals had seizures before death and thus it is likely that the large radiographic contrast load contributed to their deaths. The total contrast load needed for the angiographic studies was around 250 ml (16 ml/kg) for each animal, which is the equivalent of 1120 ml in a human adult. Previous animal experiments have demonstrated that neuronal damage caused by contrast media results from an altered blood-brain barrier. The extent of such lesions is directly proportional to the concentration and total quantity of the contrast media with a cumulative effect.9 In view of these findings we proceeded with euthanasia once seizures occurred in the remaining seven animals. In eight animals, follow-up angiograms 24 to 48 hours after embolectomy and infusion were available. Of five control limbs reexamined on follow-up, four
234
Journal of VASCULAR SURGERY
Qui~ones-Baldrich et al.
120
DISCUSSION
• Control A 8.5 c c / m i n o Experimenlol A i6.7 cc/min ~
100
0
~meon 80
*P<
+
2 S.D.
.13
U o
o,-.1
6O
o
;
ao o
0
o
40 i
8 o
I
o
.
o°
20
0
.
~
•
o
Postemb01izati0n
•
o°
Postembolectomy
Postinfusion
Fig. 5. Blood flow measurements before and after embolectomy and after intraoperative infusion. Mean increase in blood flow in experimental group suggests a trend towards increased flow when compared with controls (see text). showed no change and one was worse. Eight experimental limbs infused over 30 minutes were available for angiographic follow-up. No change from the PI angiogram was seen in two, slight but definite improvement was seen in three, worsening from the PE appearance was seen in one and return to the PE appearance was seen in two. Three experimental limbs infused over 60 minutes were available for follow-up angiographically at 48 hours. Two showed worsening and one was unchanged. Intraoperatively, there were no difficulties with hemostasis. Oozing through needle holes responded to the usual maneuvers of pressure and/or topical thrombin application. Wound exploration was carried out in 12 animals at the time of death within 24 hours after infusion and in eight animals, 24 to 48 hours after the infusion. No hematomas were found in either group. This includes five animals in which both limbs were assigned to the experimental group and thus received 120,000 units of streptokinase altogether. Three neck hematomas were present before embolectomy and fibrinolytic therapy and likely were the result of technical error. In these three animals, the hematomas remained unchanged throughout the experiment.
Stimulation of the fibrinolytic system after thromboembolectomy is an attractive approach in the management of difficult thromboembolic complications of atherosclerosis. In patients in whom therapy is delayed and in whom clot propagation has occurred, it may be the only alternative to amputation. Fear of hemorrhagic complications has kept fibrinolytic agents from being used in the operating room. There is evidence in the literature that this fear may be unwarranted. Tsapogas m in 1964 reported a case in which catheter embolectomy and intra-arterial fibrinolytic therapy were successful in reestablishing patency of a femoropopliteal graft with extensive thrombosis of the tibial vessels. There were no bleeding complications. In 1973 Feissinger et al. n reported 25 patients in whom intra-arterial urokinase was administered through a surgically placed catheter in the femoral artery. No bleeding complications were noted despite the fact that eight patients required further surgical procedures. Recently we reported our experience with five patients in whom intraoperative fibrinolytic therapy was used as an adjtmct to catheter thromboembolectomy. ~2 There were no bleeding complications. Our experimental data support the idea that regional fibrinolytic therapy is possible in an operating room setting without fear of systemic complications. Our intent was to study not only the safety, but also the effectiveness, of a short-term, relatively highdose infusion of streptokinase after embolectomy. It is clear from our data that even this short-term infusion has a positive effect in improving the angiographic result of embolectomy alone. In fact, it suggests that total blood flow is increased. Previous studies have shown that streptokinase lacks vasodilatory activity ~3 and, therefore, this increase in blood flow is likely the result of an increased vascular bed as a result of lysis of residual clot. There are several theoretical advantages to the intraoperative use offibrinolytic therapy over the percutaneous intra-arterial use. First, the bulk of material is surgically removed, thereby decreasing the amount of lysis required. Second, the infusion is carried out distal to an occluding clamp, thereby allowing for a higher concentration of the agent to develop, with plasminogen being supplied by collateral circulation. This in turn decreases the washout effect of the intact circulation, decreasing systemic effects and making it unnecessary to place a catheter adjacent to or into the thrombus. Finally, reembolectomy or other frequently required vascular procedures may be carried out when necessary during the same intervention.
Volume 4 Number 3 September 1986
Important differences between the human and canine fibrinolytic systems need to be considered. When human plasminogen is activated by a catalytic amount of streptokinase, the resulting streptokinaseplasmin complex itself carries on further activation of plasminogen. In contrast, in the canine plasminogen system, streptokinase is rapidly modified to form a modified streptokinase-plasmin complex, which is unable to carry on further activation of plasminogen. ~4 Studies on the appearance of the active center in the streptokinase-plasminogen complex revealed another important difference between human and canine plasminogen. When 1 mol of human plasminogen reacts with a stoichiometric amount of streptokinase, 1 mol of active sites is produced in 15 seconds at room temperature. In contrast, when 1 mol of canine plasminogen reacts with 1 mol ofstreptokinase, only 0.13 mol of active sites is produced in 15 seconds, and a much longer time is required for the formation of 1 mol of active centers. In the presence of high concentrations of streptokinase in the dog, activation proceeds to completion after a lag period, is In this situation, activation of canine plasminogen can be viewed as a race between the degradation of streptokinase and the activation of plasminogen. These differences suggest that fibrinolytic activity can be increased by a short-term infusion much more readily in human beings than in dogs. This in turn implies that increased effectiveness can be anticipated in the human model. The different behavior of clot and thrombus in the canine circulation as observed by us suggests increased resistance of thrombus to fibrinolysis. Some of the material that remained in the vascular tree of our animal model after embolectomy may have been in fact residual thrombus and not propagated clot. It is likely that this may account for some of the persistent defects not cleared by the infusion. O f concern is one of the experimental limbs that worsened after intraoperative fibrinolytic therapy. It was at this point that we elected to add heparin to the infusion. The thrombogenic nature of the arterial system after incomplete embolectomy and the reduction of blood flow by the proximal occluding clamp may account for this occurrence. We believe the addition of heparin to the infusion may prevent this complication. Certainly, we did not observe this problem in those infusions in which heparin was added. On the other hand, the number of experiments carried out with heparin is too small to make any definite conclusions. Fragmentation of thrombus with distal emoblization is a definite potential hazard of both percu-
Intraoperative fibrinolysis: Experimental evaluation 235
taneous and intraoperative use of fibrinolytic agents. In our animal model, defnite angiographic evidence of this complication was seen in 3 of 25 instances. Follow-up angiograms after infusion were not available in all animals. Those available were too few to provide meaningful data. However, the fact that some showed continued improvement suggests that fibrinolytic activity may remain at an elevated rate for some time after infusion. This may be the result of activated plasmin within the thrombus or residual clot that is not subject to inactivation by reestablishment of inflow. Further studies now in progress will address this particular issue. The fact that some worsened, usually returning to their status after embolectomy, suggests that postoperative anticoagulation may be essential to maintain whatever improvement was achieved by fibrinolytic therapy. From these experiments it is difficult to estimate the optimal infusion time. The number of experiments is small. In addition, such estimation with a canine model is unlikely to have clinical relevance because of aforementioned species differences. On the other hand, our results with 30-minute infusion appear superior to those with 60-minute infusion. Factors that may account for this difference include the additional time of inflow occlusion during 60minute infusion and/or inadequate dose of the agent for each unit of time. In our clinical experience, we have limited intraoperative lytic infusion time to 30 minutes. On the basis of our experimental observations, we would be hesitant to extend the infusion time without reconstitution of blood flow at timed intervals. The incidence of incomplete embolectomies after two consecutive clean passes of the Fogarty catheter in this animal experiment was 85%. Clearly, this may be due to the fragmented nature of the emboli inherent in our model. The anatomy of the arterial circulation of the hind limb in dogs may also contribute to these fragments being inaccessible to the balloon catheter. The only clinical study that has addressed this issue with angiography done after embolectomy estimated the incidence of residual clot around 40%. 7 It is likely that the true incidence in the clinical situation lies between these two numbers. These data suggest that the criterion of two consecutive clean passes of the balloon catheter is unreliable and thus angiography done after embolectomy is mandatory to ensure an adequate repair. In conclusion, we have demonstrated, with the use of a canine model of arterial thromboembolism, that (1) intra-arterial infusion of 60,000 units of streptokinase after incomplete embolectomy ira-
236
Qui~ones-Baldrich et al.
proves the angiographic appearance of the vascular bed compared with embolectomy alone; (2) the addition of heparin to the intraoperative intra-arterial infusion offibrinolytic agents is beneficial; (3) a trend toward increased blood flow is seen after the infusion of 60,000 units of streptokinase after emboleetomy; and (4) the criterion of two consecutive clean passes of the balloon catheter during embolectomy is unreliable in establishing the completeness of the repair. We acknowledge our sincere appreciation to Laura A. Jones for her assistance in the preparation of this manuscript. REFERENCES
1. Risius B, Zelch MG, Graor RA, Geisinger MA, Smith JA, Piraino DW. Catheter directed low dose streptokinase infusion: A preliminary experience. Radiology 1984; 150:34955. 2. Sussman B, Dardik H, Ibrahim IM, Fox R, Mendes D, Kahn M. Improved patient selection for enzymatic lysis of peripheral arterial and graft occlusions. Am J Surg 1984; 148:2448. 3. Van Breda A, Robinson JC, Feldman L, Waltman AC, Brewster DC, Abbott WM, Athanasoulis CA. Local thrombolysis in the treatment of arterial graft occlusions. J VAsc St:RG 1984; 1:103-12. 4. Graor RA, Risius B, Denny KM, Young JR, Beven EG, Hertzer NR, Ruschhaupt III WF, O'Hara PJ, Geisinger MA, Zelch MG. Local thrombolysis in the treatment of thrombosed arteries, bypassed grafts, and arteriovenous fistulas. J VAsc SURG 1985; 2:406-14.
Journal of VASCULAR SURGERY
5. Dunnant ]H, Edwards WS. Small vessel occlusion in the extremity after various periods of arterial obstruction: An experimental study. Surgery 1973; 240-5. 6. Greep JM, Aleman pJ, Jarret F, Bast TJ. A combined technique for peripheral arterial embolectomy. Arch Surg 1972; 105:869-74. 7. Plecha FR, Pories WJ. Intraoperative angiography in the immediate assessment of arterial reconstruction. Arch Surg 1972; 105:802-7. 8. Chandler AB. In vitro thrombotic coagulation of the blood. A method of producing a thrombus. Lab Invest 1958; 7:1104. 9. Bassett RC, Rogers JS, Cherry GR, et al. The effect of contrast media on the blood-brain barrier. J Neurosurg 1953; 10:3847. 10. Tsapogas MJ. The role of fibrinolysis in the treatment of arterial thrombosis. Experimental and clinical aspects. Ann R Coil Surg (Engl) 1964; 24:293-313. 11. Fiessinger JN, Vayssiairat M, Juillet Y, Aiach M, Janneau D, Comfier JM, Housset E, Local urokinase in arterial thromboembolism. Angiology 1980; 31:715-20. 12. Quifiones-Baldrich WJ, Zierler RE, Hiatt JC. Intraoperative fibrinolytic therapy: An adjunct to catheter thromboembolectomy. J Vasc SURG 1985; 2:319-26. 13. Tsapogas MJ, Flute IT. Experimental thrombolysis with streptokinase and urokinase. Br Med Bull 1964; 20:223-7. 14. Reddy KNN. Mechanism of activation of human plasminogen by streptokinase. In: Kline DL, Reddy KNN, eds. Fibrinolysis. Florida: CRC Press 1980:71-94. 15. Reddy KNN. Kinetics of active center formation in dog plasminogen by streptokinase and activiw of a modified streptokinase. J Biol Chem 1976; 251:6624-9.
Calif.
Percutaneous intra-arterial infusion of fibrinolytic agents has emerged as an alternative to embolectomy in selected patients with acute arterial occlusions. The combination of fibrinolytic therapy and embolectomy may be superior to either modality alone. This experiment was designed to determine safety and efficacy of intraoperative fibrinolytic therapy as an adjunct to catheter embolectomy. Forty hind limbs in 20 adult mongrel dogs were embolized with thrombus created in vitro. After 24 hours, bilateral transfemoral embolectomy was followed by an intra-arterial, intraoperative infusion. Fifteen limbs (control) received 250 ml of saline solution during a 30-minute Period; 25 limbs (experimental) received an arterial infusion of 60,000 units of streptokinase during a 30minute Period (SK 30'). In five limbs of each group, 500 units of heparin (H) was added. In five experimental limbs the streptokinase infusion time was increased to 60 minutes (SK 60'). Arteriograms and blood flow measurements were obtained before and after embolectomy (PE) and after infusion (PI); the results were compared. Improvement between the PE and PI angiograms was seen in 20% (3 of 15 dogs) of control subjects. In contrast, improvement after the infusion was evident in 100% (5 of 5 dogs) of dogs given SK plus H 30' (p < 0.01), in 80% (12 of I5 dogs) of dogs given SK 30' (p < 0.01), and in 20% (1 of 5 dogs) of dogs given SK 60'. A trend toward increased blood flow was noted in the experimental group. There were no intraoperative complications with hemostasis or postoperative bleeding (36-hour observation). We conclude that intraoperative fibrinolytic therapy in dogs is safe and effective as an adjunct to thromboembolectomy. A human clinical trial is recommended. (J VASC SURG 1986; 4:229-36.)
Catheter thromboembolectomy remains the initial procedure of choice in patients with thromboemboric complications of atherosclerosis. The intra-arterial infusion of fibrinolytic agents by percutaneous catheterization has emerged as an alternative to catheter thromboembolectomy in a selected group of patients. The results obtained with fibrinolytic therapy have been variable with partial lysis achieved in most instances and most patients requiring corrective surgery. ~-4 Experimental studies of hind limb ischemia have suggested the occurrence of arteriolar thrombosis in muscular and skin vessels after 6 hours of complete inflow occlusion/ Delays in surgical embolectomy may be hampered by similar peripheral clot propaFrom the Department of Surgery, Universityof California, Los Angeles, Center for the Health Sciences,and the ResearchDivision of the SepulvedaVeterans AdministrationMedicalCenter, Sepulveda,Calif. Supported in part by the Career DevelopmentAward Program, Veterans Administration. Reprint requests: WilliamJ. Quifiones-Baldrich,M.D., Assistant Professor of Surgery, Department of Surgery, Sectionof Vascular Surgery,UCLA Center for the Health Sciences,Los Angeles, CA 90024.
gation, which may be inaccessible to the embolectomy catheter. Greep et al.6 was able to recover additional thrombotic material in 110 peripheral arterial embolectomies with the use of a Dormia catheter in "almost all" cases after thorough Fogarty catheter embolectomy. The only clinical study that has addressed this issue with routine angiography after embolectomy reported a 36% incidence of incomplete embolectomy. 7 Therefore, incomplete removal of all intravascular clot during embolectomy is a relatively frequent problem. Bleeding complications of intra-arterial fibrinolytic therapy are a result of systemic effects of the agent, which in turn are related to the prolonged infusion time required to achieve complete resolution of the process. With these facts in mind, we postulate a role of fibrinolytic agents as an adjunct to catheter thromboembolectomy. The idea of removing the bulk of thrombus surgically and of lysing any remaining defect is attractive from the therapeutic standpoint. The purpose of these experiments is to investigate the safety and effectiveness of fibrinolytic therapy as an adjunct to catheter thromboembolectomy. 229
230 Qui~ones-Baldrich et al.
MATERIAL A N D M E T H O D S
The model of thromboembolism described in these experiments uses true thrombus. The distinction between thrombus and clot is particularly important in a canine model. Embolization of fresh clot in the canine circulation results in temporary arterial occlusion, followed by rapid lysis and dissolution because of a very active canine fibrinolytic system. Clot consists of all elements of blood, mainly red cells bound by a fibrin mesh. In contrast, thrombus consists mainly of platelets in a fbrin matrix with few red and white cells. Clot is formed by allowing coagulation to proceed in stagnant blood. Thrombus is produced when coagulation proceeds while blood is in linear motion. This concept was originally proposed by Chandler8 who described a method to produce thrombus in vitro. An aliquot of fresh blood is introduced in a 25 cm length of polyvinyl tubing and rotated in a 23-degree slanted turntable at 16 rpm. Coagulation proceeds while blood is in motion at a linear velocity of 400 cm/min. After 20 minutes, a small piece of thrombus forms at the forward tip, with the remaining blood consisting of a tail of clot or liquid from defibrinization. Using this method we have produced in vitro thrombus and embolized this material in the hind limbs of dogs. Preliminary studies in our laboratory demonstrated that embolization with thrombus produces a stable, permanent occlusion in the canine hind limb. This was compared in the same animals with clot embolization. Repeat angiography showed continued resolution with eventual complete clearing in every instance in those limbs embolized with clot. Twenty adult mongrel dogs (weighing 15 to 20 kg) of either sex underwent embolization with thrombus through a No. 7 Fr. single-lumen angiographic catheter introduced through a left carotid vessel cut down and placed under fluoroscopic guidance into the superficial femoral artery. All procedures were done with the dogs anesthetized with intravenous thiamylal (Surital), 11 mg/kg. This research was conducted under guidelines of the NIH publication "Guide for Care and Use of Laboratory Animals." Both right and left superficial femoral arteries were embolized with 4.5 to 5 ml of thrombus each. Angiography was done through the same catheter once retracted to the infrarenal aorta. An automatic film changer was programmed for two films per second for 2 seconds for the pelvis and thigh views and two films per second for 4 seconds for the distal runoff with a 3- and 4-second injection delay, respectively. Injection was achieved with a pressure
Journal of VASCULAR SURGERY
injector at 600 psi with 20 ml of 76% meglumine for the pelvis and thigh views and 30 ml of 76% meglumine for the infrageniculate runoffviews. Aortography with runoff was performed before and immediately after embolization. Twenty-four hours after embolization, transfemoral embolectomy was accomplished in all 40 limbs with the dogs placed under general endotracheal anesthesia with halothane (Fluothane) and with mechanical ventilatory support. With the use of sterile technique, both femoral arteries were exposed. Femoral blood flow was measured with an electromagnetic flowmeter (Gould Statham SP2204) before embolectomy. Three separate readings with an appropriately sized probe placed around the common femoral artery were made and recorded. Through a transverse arteriotomy in the common femoral artery, No. 3 and No. 4 Fogarty catheters were introduced distally 35 to 45 cm depending on the limb length. Embolectomy maneuvers were continued until two consecutive passes failed to retrieve any material. The amount of thrombus removed was recorded. The arterotomies were closed with interrupted 6-0 polypropylene sutures, after back-bleeding the vessels. Flow was then reestablished. Heparin (100 U/kg) was given intravenously before embolectomy. Topical bovine thrombin was used for hemostasis when necessary. Femoral blood flow measurements were taken immediately after embolectomy and recorded. Angiography was repeated immediately after completion of embolectomy. Following angiography after embolectomy, an intra-arterial infusion of either control or experimental solution was administered. A 22-gauge Teflon angiocatheter was inserted in the artery distal to an occluding vascular clamp (Fig. 1). The relationship of the arteriotomy to the occluding clamp was variable; however, the arteriotomy was always proximal to the catheter. Control solution consisted of 250 ml of normal saline solution in 10 limbs and 250 ml of normal saline solution with 500 units of heparin in five limbs. Experimental solution consisted of 250 ml of normal saline solution with 60,000 units ofstreptokinase (Streptase MSD) in 15 limbs and 250 ml of normal saline solution with 60,000 units of streptokinase and 500 units of heparin in five extremities. The infusion was given during a 30-minute period in 35 limbs (15 control and 20 experimental) and a 60-minute period in five experimental limbs (Fig. 2). Control and experimental limbs were chosen at random in the first five animals. Experimental and control limbs were chosen in the remaining 15 dogs according to results after cmbolectomy, assigning to
Volume 4 Number 3 September 1986
Intraoperative fibrinolysis: Experimental evaluation 231
/
I0- Saline
15 CONTROL
~
5- Saline + Heparin
4 0 HIND LIMBS
/
15-60,000 U SK; 50'
25 EXPERIMENTALt5 -60,000 U SK+ Heparin; 50' k - - 5 - 6 0 , 0 0 0 U SK; 60'
Fig. 2. Control and experimental groups. In 15 animals (30 limbs) one extremity was assigned to the control group and the opposite to the experimental group. In five animals (10 limbs) both extremities were assigned to the experimental group. In the latter, the right limb was infused during a 60-minute period, and the left during a 30-minute period. ~duoJ ~abus
Fig. 1. Drawing of normal arterial anatomy of canine hind limb. Note infusion site is distal to arteriotomy and an occlusive vascular clamp. the experimental group those limbs with angiographic evidence of intraluminal defects. In those animals in which both limbs had residual occlusions, the worst o f the two limbs was assigned to the experimental group. In five animals, both limbs were assigned to the experimental group, giving the intraoperative infusion after embolectomy during a 30minute period in the right limb and a 60-minute period in the left limb. Femoral blood flow measurements were done immediately after infusion and the average of three readings recorded. The angiographic catheter in the aorta was used to measure central blood pressure during blood flow measurements. An attempt to maintain the animal's blood pressure within 15 mm H g of the baseline was made. However, there were instances in which this could not be achieved and the pressure was higher or lower than baseline. Certainly this introduces an error in the absolute value o f any one particular blood flow measurement; however, in all
animals except five, one limb was control and one experimental so that the relative difference between the two sides retains validity. Angiograms were reviewed in a blinded fashion and considered without knowledge of the type of infusion administered. Specifically, changes between the angiograms performed before and after embolectomy (PE) and after infusion (PI) were noted. Improvement was considered as the reappearance of vessels not seen after embolectomy and/or disappearance of intraluminal defects causing partial restriction of flow as seen on the angiogram done after embolectomy. If no change was seen between these two studies, the result was recorded as unchanged. If the PI angiogram failed to show vessels that were seen on the angiogram after embolectomy, the result was recorded as worsened. When possible, angiography was repeated between 24 and 48 hours after the infusion. However, the total contrast load given to these animals precluded survival for this length of time. In future studies, we would modify the protocol and eliminate both the baseline and the angiogram after embolization, because the former was useful in only one instance and the latter did not add any valuable information. RESULTS The baseline angiogram revealed normal anatomy in 38 of 40 limbs. Two limbs in the same animal showed the popliteal artery ending in small branches below the knee and the distal circulation supplied by the saphenous artery. Embolization with thrombus achieved an occlusion of the superficial femoral artery in all instances. Transfemoral embolectomy retrieved thrombus and clot in all instances, the average recovered being 3 + 1 ml. The Fogarty balloon catheter could be introduced for the length of the extremity in 38 limbs. In two limbs, an anomalous popliteal artery as seen in the baseline angiogram
Journal of VASCULAR SURGERY
232 Qui~ones-Baldrich et M.
Fig. 3. A, Angiogram after bilateral transfemoral embolectomies shows saphenous artery and distal tibial artery occlusion in right limb (curped and lade stra~qht arrow). Left system appears intact except for some defects in thigh branches. B, Angiogram after experimental 30-minute infusion (no heparin) in the right limb and control infusion in the left limb. Note significant improvement in both saphenous and tibial arteries on right. The left (control) distal tibial artery is now occluded with reconstitution of the pedal arch by collateral circulation. precluded passage of the catheter beyond the infrageniculate region. Residual clot and thrombus after embolectomy was seen in 34 of 40 (85%) limbs on angiograms done after embolectomy. Embolectomy was complete in six limbs (four control and two experimental). Improvement between the PE and PI angiograms was seen in 2 of i0 control limbs infused with 250 ml of saline solution. In this group, seven limbs were unchanged and one was worse after infusion (Fig. 3). One of five control limbs infused with saline solution and heparin appeared improved, whereas one was worse and three unchanged after infusion. All control limbs were infused over 30 minutes. Thus, a total of 3 of 15 control limbs showed improvement between the PE and PI angiograms (Fig. 4) with no difference seen between the group that received saline solution alone and that which received saline solution
plus heparin. In contrast, improvement between the PE and PI angiograms was seen in 12 of 15 limbs infused over 30 minutes with 60,000 units of streptokinase dissolved in 250 ml of saline solution. In two limbs there was no appreciable difference between the PE and PI angiograms. In one limb the PI angiogram was worse, showing reocclusion of the superficial femoral artery. The difference between the control (3 of 15 dogs) and experimental (12 of 15 dogs) groups is statistically significant (p < 0.003) by Fisher's exact two-tailed test. Five experimental limbs were infused with a solution of 60,000 units ofstreptokinase and 500 units of heparin in 250 ml of normal saline solution. PI angiograms revealed improvement in all five extremities (Fig. 4). This is a statistically significant difference when compared with the control group (p < 0.004); however, it is not significant when
Volume 4 Number 3 September 1986
Intraoperative fibrinolysis: Experimental evaluation 233
Fig. 4. A, Angiogram done after embolectomy shows residual clot bilaterally (arrows). B, After control infusion with heparin in right limb residual clot is unchanged, with some improvement in the middle digital artery. After 60,000 units of streptokinase with 1000 units of heparin was infused there is remarkable improvement in all three digital arteries with possible migration of intravascular clot (left arrow). Both results were recorded as improved. compared with the experimental group receiving no intra-arterial heparin in conjunction with the streptokinase infusion. In five experimental limbs, the infusion time was 60 minutes, the infusate consisting of 60,000 units of streptokinase in 250 ml of normal saline solution. Improvement was seen in one of the five limbs. In the remaining four limbs no appreciable differences between the PE and the PI angiograms were seen. In these animals, the opposite limb was infused over 30 minutes with 60,000 units of streptokinase in 250 ml of normal saline solution. Angiography showed improvement in three limbs, one was unchanged, and in one worsening of the PI appearance was evident. The 30-minute results are included in the first experimental group already presented. The results of blood flow measurements before and after embolectomy and after the intraoperative intra-arterial infusion are shown in Fig. 5. The mean + 2 standard deviations of the blood flow after embolectomy was 41.1 _ 6.4 ml/min for control and 30.7 _+ 5.0 ml/min for experimental limbs. This lower figure in the experimental limbs is the result of selecting the worst of the embolized limbs for the experimental group. The mean blood flow after infusion was 49.9 _+ 6.4 ml/min for control and 47.4 + 6.3 ml/min for experimental limbs. The mean increase (A) of 8.3 ml/min in the control group
vs. 16.7 ml/min in the experimental group does not achieve statistical significance (p < 0.13). However, this is likely the result of the small number of experiments and suggests a positive trend toward increased blood flow in the experimental group. When possible, angiography was repeated 24 to 48 hours after infusion. Twelve animals died within the 24 hours after embolectomy and infusion. Complete autopsy was performed in the first five deaths; no abnormalities were found to explain the outcome. Specifically, no intra-abdominal or intrathoracic bleeding was present. Most of these animals had seizures before death and thus it is likely that the large radiographic contrast load contributed to their deaths. The total contrast load needed for the angiographic studies was around 250 ml (16 ml/kg) for each animal, which is the equivalent of 1120 ml in a human adult. Previous animal experiments have demonstrated that neuronal damage caused by contrast media results from an altered blood-brain barrier. The extent of such lesions is directly proportional to the concentration and total quantity of the contrast media with a cumulative effect.9 In view of these findings we proceeded with euthanasia once seizures occurred in the remaining seven animals. In eight animals, follow-up angiograms 24 to 48 hours after embolectomy and infusion were available. Of five control limbs reexamined on follow-up, four
234
Journal of VASCULAR SURGERY
Qui~ones-Baldrich et al.
120
DISCUSSION
• Control A 8.5 c c / m i n o Experimenlol A i6.7 cc/min ~
100
0
~meon 80
*P<
+
2 S.D.
.13
U o
o,-.1
6O
o
;
ao o
0
o
40 i
8 o
I
o
.
o°
20
0
.
~
•
o
Postemb01izati0n
•
o°
Postembolectomy
Postinfusion
Fig. 5. Blood flow measurements before and after embolectomy and after intraoperative infusion. Mean increase in blood flow in experimental group suggests a trend towards increased flow when compared with controls (see text). showed no change and one was worse. Eight experimental limbs infused over 30 minutes were available for angiographic follow-up. No change from the PI angiogram was seen in two, slight but definite improvement was seen in three, worsening from the PE appearance was seen in one and return to the PE appearance was seen in two. Three experimental limbs infused over 60 minutes were available for follow-up angiographically at 48 hours. Two showed worsening and one was unchanged. Intraoperatively, there were no difficulties with hemostasis. Oozing through needle holes responded to the usual maneuvers of pressure and/or topical thrombin application. Wound exploration was carried out in 12 animals at the time of death within 24 hours after infusion and in eight animals, 24 to 48 hours after the infusion. No hematomas were found in either group. This includes five animals in which both limbs were assigned to the experimental group and thus received 120,000 units of streptokinase altogether. Three neck hematomas were present before embolectomy and fibrinolytic therapy and likely were the result of technical error. In these three animals, the hematomas remained unchanged throughout the experiment.
Stimulation of the fibrinolytic system after thromboembolectomy is an attractive approach in the management of difficult thromboembolic complications of atherosclerosis. In patients in whom therapy is delayed and in whom clot propagation has occurred, it may be the only alternative to amputation. Fear of hemorrhagic complications has kept fibrinolytic agents from being used in the operating room. There is evidence in the literature that this fear may be unwarranted. Tsapogas m in 1964 reported a case in which catheter embolectomy and intra-arterial fibrinolytic therapy were successful in reestablishing patency of a femoropopliteal graft with extensive thrombosis of the tibial vessels. There were no bleeding complications. In 1973 Feissinger et al. n reported 25 patients in whom intra-arterial urokinase was administered through a surgically placed catheter in the femoral artery. No bleeding complications were noted despite the fact that eight patients required further surgical procedures. Recently we reported our experience with five patients in whom intraoperative fibrinolytic therapy was used as an adjtmct to catheter thromboembolectomy. ~2 There were no bleeding complications. Our experimental data support the idea that regional fibrinolytic therapy is possible in an operating room setting without fear of systemic complications. Our intent was to study not only the safety, but also the effectiveness, of a short-term, relatively highdose infusion of streptokinase after embolectomy. It is clear from our data that even this short-term infusion has a positive effect in improving the angiographic result of embolectomy alone. In fact, it suggests that total blood flow is increased. Previous studies have shown that streptokinase lacks vasodilatory activity ~3 and, therefore, this increase in blood flow is likely the result of an increased vascular bed as a result of lysis of residual clot. There are several theoretical advantages to the intraoperative use offibrinolytic therapy over the percutaneous intra-arterial use. First, the bulk of material is surgically removed, thereby decreasing the amount of lysis required. Second, the infusion is carried out distal to an occluding clamp, thereby allowing for a higher concentration of the agent to develop, with plasminogen being supplied by collateral circulation. This in turn decreases the washout effect of the intact circulation, decreasing systemic effects and making it unnecessary to place a catheter adjacent to or into the thrombus. Finally, reembolectomy or other frequently required vascular procedures may be carried out when necessary during the same intervention.
Volume 4 Number 3 September 1986
Important differences between the human and canine fibrinolytic systems need to be considered. When human plasminogen is activated by a catalytic amount of streptokinase, the resulting streptokinaseplasmin complex itself carries on further activation of plasminogen. In contrast, in the canine plasminogen system, streptokinase is rapidly modified to form a modified streptokinase-plasmin complex, which is unable to carry on further activation of plasminogen. ~4 Studies on the appearance of the active center in the streptokinase-plasminogen complex revealed another important difference between human and canine plasminogen. When 1 mol of human plasminogen reacts with a stoichiometric amount of streptokinase, 1 mol of active sites is produced in 15 seconds at room temperature. In contrast, when 1 mol of canine plasminogen reacts with 1 mol ofstreptokinase, only 0.13 mol of active sites is produced in 15 seconds, and a much longer time is required for the formation of 1 mol of active centers. In the presence of high concentrations of streptokinase in the dog, activation proceeds to completion after a lag period, is In this situation, activation of canine plasminogen can be viewed as a race between the degradation of streptokinase and the activation of plasminogen. These differences suggest that fibrinolytic activity can be increased by a short-term infusion much more readily in human beings than in dogs. This in turn implies that increased effectiveness can be anticipated in the human model. The different behavior of clot and thrombus in the canine circulation as observed by us suggests increased resistance of thrombus to fibrinolysis. Some of the material that remained in the vascular tree of our animal model after embolectomy may have been in fact residual thrombus and not propagated clot. It is likely that this may account for some of the persistent defects not cleared by the infusion. O f concern is one of the experimental limbs that worsened after intraoperative fibrinolytic therapy. It was at this point that we elected to add heparin to the infusion. The thrombogenic nature of the arterial system after incomplete embolectomy and the reduction of blood flow by the proximal occluding clamp may account for this occurrence. We believe the addition of heparin to the infusion may prevent this complication. Certainly, we did not observe this problem in those infusions in which heparin was added. On the other hand, the number of experiments carried out with heparin is too small to make any definite conclusions. Fragmentation of thrombus with distal emoblization is a definite potential hazard of both percu-
Intraoperative fibrinolysis: Experimental evaluation 235
taneous and intraoperative use of fibrinolytic agents. In our animal model, defnite angiographic evidence of this complication was seen in 3 of 25 instances. Follow-up angiograms after infusion were not available in all animals. Those available were too few to provide meaningful data. However, the fact that some showed continued improvement suggests that fibrinolytic activity may remain at an elevated rate for some time after infusion. This may be the result of activated plasmin within the thrombus or residual clot that is not subject to inactivation by reestablishment of inflow. Further studies now in progress will address this particular issue. The fact that some worsened, usually returning to their status after embolectomy, suggests that postoperative anticoagulation may be essential to maintain whatever improvement was achieved by fibrinolytic therapy. From these experiments it is difficult to estimate the optimal infusion time. The number of experiments is small. In addition, such estimation with a canine model is unlikely to have clinical relevance because of aforementioned species differences. On the other hand, our results with 30-minute infusion appear superior to those with 60-minute infusion. Factors that may account for this difference include the additional time of inflow occlusion during 60minute infusion and/or inadequate dose of the agent for each unit of time. In our clinical experience, we have limited intraoperative lytic infusion time to 30 minutes. On the basis of our experimental observations, we would be hesitant to extend the infusion time without reconstitution of blood flow at timed intervals. The incidence of incomplete embolectomies after two consecutive clean passes of the Fogarty catheter in this animal experiment was 85%. Clearly, this may be due to the fragmented nature of the emboli inherent in our model. The anatomy of the arterial circulation of the hind limb in dogs may also contribute to these fragments being inaccessible to the balloon catheter. The only clinical study that has addressed this issue with angiography done after embolectomy estimated the incidence of residual clot around 40%. 7 It is likely that the true incidence in the clinical situation lies between these two numbers. These data suggest that the criterion of two consecutive clean passes of the balloon catheter is unreliable and thus angiography done after embolectomy is mandatory to ensure an adequate repair. In conclusion, we have demonstrated, with the use of a canine model of arterial thromboembolism, that (1) intra-arterial infusion of 60,000 units of streptokinase after incomplete embolectomy ira-
236
Qui~ones-Baldrich et al.
proves the angiographic appearance of the vascular bed compared with embolectomy alone; (2) the addition of heparin to the intraoperative intra-arterial infusion offibrinolytic agents is beneficial; (3) a trend toward increased blood flow is seen after the infusion of 60,000 units of streptokinase after emboleetomy; and (4) the criterion of two consecutive clean passes of the balloon catheter during embolectomy is unreliable in establishing the completeness of the repair. We acknowledge our sincere appreciation to Laura A. Jones for her assistance in the preparation of this manuscript. REFERENCES
1. Risius B, Zelch MG, Graor RA, Geisinger MA, Smith JA, Piraino DW. Catheter directed low dose streptokinase infusion: A preliminary experience. Radiology 1984; 150:34955. 2. Sussman B, Dardik H, Ibrahim IM, Fox R, Mendes D, Kahn M. Improved patient selection for enzymatic lysis of peripheral arterial and graft occlusions. Am J Surg 1984; 148:2448. 3. Van Breda A, Robinson JC, Feldman L, Waltman AC, Brewster DC, Abbott WM, Athanasoulis CA. Local thrombolysis in the treatment of arterial graft occlusions. J VAsc St:RG 1984; 1:103-12. 4. Graor RA, Risius B, Denny KM, Young JR, Beven EG, Hertzer NR, Ruschhaupt III WF, O'Hara PJ, Geisinger MA, Zelch MG. Local thrombolysis in the treatment of thrombosed arteries, bypassed grafts, and arteriovenous fistulas. J VAsc SURG 1985; 2:406-14.
Journal of VASCULAR SURGERY
5. Dunnant ]H, Edwards WS. Small vessel occlusion in the extremity after various periods of arterial obstruction: An experimental study. Surgery 1973; 240-5. 6. Greep JM, Aleman pJ, Jarret F, Bast TJ. A combined technique for peripheral arterial embolectomy. Arch Surg 1972; 105:869-74. 7. Plecha FR, Pories WJ. Intraoperative angiography in the immediate assessment of arterial reconstruction. Arch Surg 1972; 105:802-7. 8. Chandler AB. In vitro thrombotic coagulation of the blood. A method of producing a thrombus. Lab Invest 1958; 7:1104. 9. Bassett RC, Rogers JS, Cherry GR, et al. The effect of contrast media on the blood-brain barrier. J Neurosurg 1953; 10:3847. 10. Tsapogas MJ. The role of fibrinolysis in the treatment of arterial thrombosis. Experimental and clinical aspects. Ann R Coil Surg (Engl) 1964; 24:293-313. 11. Fiessinger JN, Vayssiairat M, Juillet Y, Aiach M, Janneau D, Comfier JM, Housset E, Local urokinase in arterial thromboembolism. Angiology 1980; 31:715-20. 12. Quifiones-Baldrich WJ, Zierler RE, Hiatt JC. Intraoperative fibrinolytic therapy: An adjunct to catheter thromboembolectomy. J Vasc SURG 1985; 2:319-26. 13. Tsapogas MJ, Flute IT. Experimental thrombolysis with streptokinase and urokinase. Br Med Bull 1964; 20:223-7. 14. Reddy KNN. Mechanism of activation of human plasminogen by streptokinase. In: Kline DL, Reddy KNN, eds. Fibrinolysis. Florida: CRC Press 1980:71-94. 15. Reddy KNN. Kinetics of active center formation in dog plasminogen by streptokinase and activiw of a modified streptokinase. J Biol Chem 1976; 251:6624-9.