- Email: [email protected]
Failure of thrombolytic therapy to improve long-term vascular patency
Failure of thrombolytic therapy to improve long-term vascular patency Gian Luca Faggioli, MD, Richard M. Peer, MD, Luciano Pedrini, MD, Marco Donato Di Paola, MD, James A. Upson, MD, Massimo D'Addato, MD, and John J. Ricotta, MD, Buffalo) N.Y,) and Bologna) Italy Purpose: Few data are available on long-term follow-up of arterial segments subjected to thrombolysis. We reviewed all cases of vascular occlusion treated with urokinase to identify early success and determine the influence of postlysis intervention and the nature of the thrombosed segment (i.e., artery vs graft) on long-term patency. Methods: Data on 134 cases (58 arteries, 76 grafts) treated with high-dose urokinase infusion in the lower limbs over a 7 -year period were analyzed. Limbs were divided into five groups on the basis of therapy after lytic infusion to determine long-term efficacy: group I, success with no additional therapy; group II, percutaneous angioplasty alone; group III, limited surgical procedure (operative angioplasty, jump graft); group IV, extensive procedure (new bypass); and group V, revascularization after lytic failure. Long-term results were assessed by life-table analysis and groups compared by log-rank test (Mantel-Haenszel). Results: Initial patency was established in 87 (64.9%) of 134 cases with 5 deaths (3.7%), 11 amputations (8.2%), and 16 complications (11.9%). Follow-up was available in 68.6% of cases for a mean of 10.9 months. No difference was seen between grafts and native arteries. Patency was analyzed at 6, 12, 18, and 24 months. The 24-month patency rate after lysis alone (group I - 25.9%) was inferior (p < 0.05) to results after lysis and any subsequent intervention (groups II, III, and IV). The type of intervention did not influence subsequent patency. Twenty-four-month patency of procedures performed after failed thrombolysis (group V, 41.4%) was not different from those after successful lysis (groups I to IV). Twenty-four-month patency in groups II and III (minor interventions, 62.9%) was not significantly different from that of groups IV and V (major interventions, 53.2%) (p > 0.25). Conclusions: Operative intervention is required to produce long-term arterial patency, even after successful thrombolysis. No statistically significant benefit of thrombolysis on vascular patency was seen in our series. (J VAse SURG 1994;19:289-97.)
Intrarterial thrombolytic therapy with urokinase is commonly used for the treatment of acute and subacute thrombosis of arteries and grafts. The rationale for this approach is based on good early results, acceptable complication rates, the possibility of minimizing subsequent interventions, and the theoretical capacity to dissolve clots in the smaller peripheral vessels. l . S There are few data in the From the Department of Surgery, Division of Vascular Surgery, State University of New York at Buffalo, and Department of Vascular Surgery (Drs. Pedrini, Di Paola, and D'Addato), University of Bologna. Presented at rhe Forty-sevenrh Annual Meeting of rhe Society for Vascular Surgery, Washington, D.C., June 8-9, 1993. Reprint requests: John J. Ricotta, MD, Department of Surgery, Division of Vascular Surgery, Millard Fillmore Hospital, 3 Gates Circle, Buffalo, NY 14209. Copyright © 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, Norrh American Chapter. 0741-S214(94($3.00 + 0 24/6/50763
literature that address the long-term fate of vessels successfully recanalized by thrombolytic infusion, with or without the use of adjunctive procedures. l ,2,4.8 Without these data the long-term benefit of thrombolysis as a therapeutic intervention will be difficult to assess. To obtain some of this information, we examined the long-term results obtained in a population of patients treated with intraarterial urokinase therapy for lower limb arterial occlusions.
MATERIAL AND METHODS The records of all thrombolytic infusions performed for lower limb arterial or graft thrombosis over a 7-year period were reviewed. All patients were diagnosed with class I and II ischemia with Society for Vascular Surgery/International Society for Cardiovascular Surgery, North American Chapter criterion. 9 Urokinase was used as thrombolytic agent and 289
JOURNAL OF VASCULAR SURGERY February 1994
290 Faggioli et at.
Table 1. Early results of thrombolytic therapy in arteries and grafts No.
Arteries Suprainguinal Infrainguinal All Grafts Suprainguinal Infrainguinal All Total Suprainguinal Infrainguinal All
Failures
Complication
Death
4 (21.0%) 15 (38.5%) 19 (32.7%)
4 (21%) 10 (25.6%) 14 (24.1%)
1 (2.5%)
2 (8%) 11 (21.5%) 13 (17.1%)
1 (1.9%)
9 (11.9%)
7 (28%) 21 (41.2%) 28 (36.8%)
5 (11.4%) (12.3%) 16 (11.9%)
11 (25%) 36 (40%) 47 (35.1%)
6 (13.6%) 21 (23.3%) 27 (20%)
3 (6.8%) 2 (2.2%) 5' (3.7%)
CL
PL
19 39 58
12 (63.2%) 20 (51.3%) 32 (55.2%)
± (10.2%)
25 51 76
16 (64%) 23 (45.1%) 39 (51.3%)
Z (13.8%)
44 90 134
28 (63.6%) 43 (47.7%) 71 (52.9%)
3 (15.8%)
7 (12.1%)
2 (8%)
11
0(0%)
1 (1.7%) 3 (12%)
4 (5.2%)
P > 0.05 between rows.
delivered intraarterially with the tip of the catheter inserted into the thrombus whenever possible. Over the course of time, different infusion protocols were followed. These include the protocol described by McNamara and FischerlO (initial dose of 4000 IV/min followed by a maintenance dose of 1000 IV/min), and other protocols described in the literature (75,000 to 150,000 IV bolus dose with 25,000 to 75,000 IV/hr maintenance dose).)),12 Concomitant intravenous heparin infusion was used in most cases. Angiograms were obtained before, during, and at completion of the thrombolytic treatment. Results of the lytic infusions were classified as complete lysis (CL, restoration of flow throughout the vessel with no evidence of residual clot), partial lysis (PL, achievement of patency with minimal residual clot inside the lumen), and failure (absence of or minimal clot lysis). The patient population was then divided in five groups according to the result of the lytic treatment and the type of subsequent treatment performed for analysis of long-term results: group I, success of thrombolytic therapy (CL or PL) with no evidence of correctable lesions and no additional therapy; group II, success of thrombolytic therapy with minor defects in the lumen corrected by percutaneous trans luminal angioplasty (PTA); group III, success of thrombolytic therapy with a localized residual defect (a limited surgical procedure [i.e., angioplasty, jump graft] was performed); group IV, success of thrombolytic therapy with extensive residual disease of the vessel or the graft (an extensive surgical procedure [i.e., new bypass, extensive proximal or distal reconstruction] was required in these cases]; group V, failure of lytic therapy and surgical revascularization. Follow-up was obtained by direct examination and vascular laboratory evaluation of the patient. Immediate patency rates were compared by chi-
squared analysis. Differences between late results were analyzed by life-table analysis (Kaplan-Meier) and compared by log-rank test (Mantel-Haenszel). Statistical significance was defined as p < 0.05.
RESULTS Complete information on 134 urokinase infusions in 122 patients was available. There were 76 (56.7%) grafts (25 suprainguinal and 51 infrainguinal) and 58 (43.3%) native arteries (19 suprainguinal and 39 infrainguinal). The mean duration of symptoms was 4.6 ± 8.8 days (median 1.5), and the mean infusion time was 53.8 hours ± 29.0 (median 48). Early results Complete lysis was obtained in 71 cases (52.9%) and partial lysis in 16 (11.9% ), for a total of 8 7 of 134 (64.9%) successful cases. A complete description of early results in grafts and arteries according to their location is reported in Table I. No statistical difference in efficacy or mortality and morbidity rates was found by chi-squared analysis between graft and native artery or between infrainguinal and suprainguinallocation (p > 0.05) (Table I). Five deaths occurred for an overall mortality rate of 3.7%. The causes of death were directly related to the infusion therapy and included cerebral hemorrhage (three cases), myocardial infarction (one case), and gastrointestinal bleeding (one case). Major complications occurred in 27 cases (20.1%): these included 11 amputations as a result offailed thrombolytic therapy and prolonged ischemia, 10 hematomas that required surgical evacuation or blood transfusion, two peripheral embolizations, one case of pulmonary edema, one myocardial infarction, one lacunar infarction, and one catheter thrombosis. Mortality and morbidity rates were evenly distributed
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
Paggioli et al.
291
Table II. Early complications and long-term patency (± SE) in treatment groups Complications*
Group I
Perioperative
6 mas
12 mas
18 mas
24 mas
25.90/0 ± 21.0%
1 (3%)
96.9% ± 3.3%
51.0% ± 9.5%
39.3% ± 12.4%
25.9% ± 12.5%
0
77.8% ± 12.3%
77.8% ± 18.5%
58.3% ± 26.7%
58.3% ± 37.6%t
3 (13.6%)
81.8% ± 7.5%
63.0% ± 11.5%
63.0% ± 14.4%
63.0% ± 19.1%t
63.0% ± 27.0%t
3 (13.0%)
91.3% ± 5.7%
78.2% ± 10.1%
56.3% ± 13.1%
56.3% ± 16.6%t
56.3% ± 21.4%t
3 (14.3%)
80.0% ± 7.8%
57.2% ± 11.8%:j:
51.8% ± 13.6%:j:
41.4% ± 14.1%:j:
41.4% ± 18.1%:j:
(n = 33)
Group II (n = 9)
Group III (n = 22) Group IV (n = 23) Group V (n = 21)
*Group I, myocardial infarction; group III, one embolization and two hematomas; group IV, one embolization and two hematomas; and group V, three hematomas. tp < 0.05 compared with group 1. :j:p = NS compared with other groups.
in the different graft types and anatomic categories. Previous investigators have correlated duration of infusion (> 12 to 24 hours) with systemic lytic effect. 13 In our patients the duration of treatment, that is, less or more than 18 hours, did not significantly influence incidence of death or disease. The time from the occlusive event to the beginning of lytic treatment was recorded in 101 cases. Fourteen patients (13.8%) were treated after more than 7 days of occlusion, with CL in six (42.8%) patients, PL in two (14.3%), and failures in six (42.8%) patients. In the 87 patients (86.1%) treated within a week from occlusion there was CL in 38 (55.2%) patients, PL in 12 (13.8%), and failures in 27 (31%) patients. The difference between these groups (8 ofl4 = 57.1%; 50 of87 = 57.5%) is not statistically significant (p > 0.25). However, the morbidity and mortality rates in the group with longer duration of occlusion were significantly greater (71.4% = five complications, four amputations, one death) than the rates of the group with an occlusion time less than 7 days (17.2% = five complications, SIX amputations, four deaths) (p < 0.001). Long-term results The distribution of cases by result of lytic therapy was as follows. There were 33 cases (24.6%) in group I, in which no additional procedures were performed because of the inability to detect a correctable lesion. Twenty-seven (81.9%) of these cases showed complete lysis and six (18.1 %) had some minor residual thrombi. All patients in this group underwent anticoagulant therapy. The presence of residual clot in these six cases did not seem to influence subsequent follow-up, because only one case with PL occluded within 1 month and the others remained patent for at
least 6 months. The residual stenosis after lytic therapy was treated by PTA in nine cases (6.7%, group II). Twenty-two cases (16.4%) were treated by "minor" surgical procedures, including 10 jump grafts, 10 surgical angioplasties, and two thromboembolectomies (group III). In 23 other cases (17.2%) an "extensive" surgical procedure was performed (group IV): in seven cases this was an aortobifemoral bypass and in 16 cases a new femoropopliteal or femorodistal bypass. Twenty-one cases ( 15 .7% ) in which the lytic therapy failed were treated surgically (group V); the intervention performed was a femoropopliteal or femorodistal bypass in 10 cases, a thromboembolectomy in five cases, a thromboembolectomy with angioplasty in three cases, and an axillofemoral bypass in three cases. There was no difference in perioperative complications between any of the treatment groups (Table II). Twenty-six cases (19.4%) are not included in the five groups listed above. They consist of the five deaths and the 11 primary amputations that occurred during lytic treatment and 10 other patients who had clinical signs of improvement in the ischemic limb despite unsuccessful thrombolytic treatment. Three of these patients were lost to follow-up, three were admitted again with rest pain at 7, 18, and 36 months; and four patients had limiting claudication at 5, 6, 16, and 18 months. Overall, thrombolysis was effective in avoiding major vascular reconstruction in 64 of 134 cases (47.8%). Follow-up was available in 74 of 108 (68.6%) cases included in the five treatment groups, for a mean of 10.9 ± 10.8 months (range 1 to 60 months). Life-table analyses of each group of treatment are reported in Table II and in Fig. 1). Results in group I were significantly inferior to those of the other groups of intervention after successful throm-
JOURNAL OF VASCULAR SURGERY February 1994
292 Faggioli et al.
100
80
-- GROUP I
.., > U Z
I
-0-
.•.•.-t ............................... ..
60
L4-...j.QooQ-ooQool~i-'---..g.... - - -
w
GROUP II
.... GROUP III
- -.
~
-- GROUP IV 40
-- GROUP V
20
23 5 14 17 15
19 5 13 16 14
14 4 11 13 10
9 4 11 11 10
8
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
0 0
2 11
6 2 7
6 1 4
3 1 4
8
8
8
8
7
7
7
5
4 5 5
2 3 3
2 3 3
2 3 3
2
2
GI Gil Gill GIV GV
FOLLOW-UP (Months) Fig. 1. Life-table analyses in five treatment groups (group I, success with no further therapy; group II, percutaneous angioplasty alone; group III, limited surgical procedure; group IV, extensive procedure; and group V, revascularization after lytic failure).
bolysis (groups II, III, and IV). This was true both at the 18-month interval (group I vs group II, p = 0.01; group I vs group III or IV,p < 0.05) and at the 24-month interval (group I vs group III or IV, p < 0.05). When surgical intervention was required, long-term patency was similar regardless of the extent of the procedure used (group III vs group IV, p < 0.25 at 18 and 24 months). Patients who underwent surgical reconstruction after failure of thrombolysis (group V) demonstrated a reduced long-term patency (41% at 24 months), although differences with the other treatment groups were not statistically significant at 18 and 24 months. Patency rates in patients who underwent "minor" treatment after thrombolysis (groups II and III = 62.9% at 18 and 24 months) were not significantly different from those of patients in whom an "extensive" intervention was required, either after successful or unsuccessful thrombolysis (groups IV and V = 53.2% at 18 and 24 months; p > 0.25) (Fig. 2). Regardless of the treatment performed after the lytic therapy, grafts and arteries demonstrated similar long-term patency rates (58.7% ± 12.5% vs 56.6% ± 18.6% at 18 months,p > 0.25). In native arteries, cases in groups II and III ("minor" inter-
vention) had better long-term results than those in groups IV and V ("extensive" treatment) (100% vs 55.8% at 18 months), although this approached but did not reach statistical significance (p = 0.06). Native arteries also showed better patency rates than grafts in groups II and III (100% vs 55.8%, P = 0.06) (Table III). Suprainguinal grafts tended to behave better than infrainguinal grafts after a 12-month interval (p = 0.06), but this effect was lost at 18 and 24 months (p = 0.1). No significant difference was found when 18- and 24-month patency rates of suprainguinal arteries were compared with patency rates of infrainguinal arteries (p > 0.1). However, the long-term patency rates in groups II and III was significantly greater in suprainguinal (91.6% at 18 months) than in infrainguinal grafts (39.6%; p < 0.025). In suprainguinal grafts groups II and III also had better 18-month patency rates than groups IV and V (32.7%;p < 0.05) (Table III). Saphenous veins had the lowest 6- and 12-month patency rates (38.3% ± 11.3% and 25.5% ± 15.5%), but these results were not significantly different from that of polytetrafluoroethylene grafts (64.1 % ± 15.7% and 51.3% ± 20.6%;p > 0.1).
JOURNAL OF VASCULAR SURGERY Volwne 19, Number 2
Faggioli et al.
293
100
80
-- GROUP I
:>
u 60
GROUP II & III
Z
-0-
~
.... GROUP IV & V
w
Q.
40
20
0
0
23 19 32
19 14 18 15 30 23
2
4
6
9 8 6 15 13 9 21 15 15
6
5 15
1 3 5 4 13 10
1 2
1 2
1 2
6
6
6
2
2
GI GII&III GIV&V
8 10 12 14 16 18 20 22 24 26 28 30
FOLLOW-UP (Months) Fig. 2. Comparison oflife-table analyses in group I (success with no further therapy); groups II and III ("minor" procedure after successful thrombolysis); and groups IV and V ("extensive" procedure after successful or failed thrombolysis).
DISCUSSION Intraarterial, preferably intrathrombus, urokinase infusion has been proposed as the first-choice treatment for acute arterial and graft thrombosis. Immediate success rates with intrathrombus urokinase can be as high as 88% in both arteries and grafts,3-8 and a metaanalysis of the English literature reveals an average success rate of85% in the most recent series, with a complication rate of 20%.14 There are many potential attractions of thrombolytic therapy over an initial surgical approach. They include the ability to restore patency with a percutaneous procedure, the possibility of improved patency of run-off vessels with the increased ability to plan subsequent intervention, and the potential to simplify subsequent surgical procedures. 1-5,15-18 Some authors state that the very distal arterial bed can be cleared of small thrombi more efficiently than with balloon embolectomy, without jeopardizing the endothelial integrity.l,2 In aggregate, these potential advantages have been proposed to justify thrombolysis as the procedure of choice and are believed by many to offset its complication rate and cost. 4 ,5,15-19 Acceptance of thrombolysis as the procedure of choice has not been universal. Skeptics point out the
Table III. Long-term results in groups II and III ("minor" procedures) and groups IV and V ("extensive" procedures)
Infrainguinal Suprainguinal Arteries Grafts <7 days ocelusion >7 days ocelusion
*p
=
Groups II andlli
Groups IV and V
p Value
39.6%* 91.6%* 100%* 55.8%* 58.0%
62.4 32.7% 45.8% 60.0% 48.1%
NS <0.05 <0.05 NS NS
60.0%
30.0%
NS
0.06 within treatment groups.
cost and complications of the technique and frequent need for adjunctive surgical procedures. 2o-28 No prospective randomized data are available. In a series of 27 cases treated with streptokinase, Wolfson, Kumpe, and Rutherford7 found an I8-month patency rate of 22% and 50% in grafts and arteries, respectively. Graor et al. 1 retrospectively compared 33 patients who underwent thrombolysis with 38 cases treated surgically. Thirty-day patency rate was significantly higher in the lytic treatment group (86%
294 Faggioli et al.
vs 42%), but the I-year patency rates were 20% and 14%, a difference that is unlikely to be significant in such a small population. 1 McNamara and Bomberger6 found a 6-month patency rate of 59% ina group of 100 cases treated with high-dose urokinase; patency rates of vein grafts after successful thrombolysis were 36.6% and 22.9% at 1 and 2 years in the study from Belkin et aI.2 The most favorable results were reported by Sullivan et al.,8 and were as high as 79% patency at 2 years in the cases with a correctable lesion after successful lytic treatment. If thrombolysis cannot be shown to improve long-term vessel patency and limb salvage, its routine use with attendant risks and expense will be difficult to justify. Therefore an evaluation oflong-term patency after lytic infusion is critical. There must be clear benefit associated with thrombolytic infusion to offset its cost and complication rate. Infusion must either reduce short-term morbidity rates, mortality rates, and cost without compromising long-term patency or provide improved long-term patency without significant increase in short-term complications. Our study was designed to retrospectively review the short-term complications and long-term patency associated with the use of thrombolytic infusion. Whereas analysis of these data is limited by the constraints of a retrospective study, it can provide insight about conditions in which thrombolysis is of potential benefit. Our practice has been to use thrombolytic infusion in all the patients with acute thrombotic occlusion, in the absence of specific contraindications. Despite different protocols for urokinase delivery, we could find no relationship of infusion protocols to complication rate or results in our series. The retrospective nature of our review and variety of doses used precludes any conclusion in this regard. We did not see increased morbidity rates associated with operative intervention. Although our success rate was somewhat lower than that reported by some authors, it is comparable to other experiences in the literature. 2 ,16,25 Thrombolysis was successful in 65% of cases, associated with a 3.7% mortality rate and a 20% complication rate. Although about one third of the complications were amputations (S.2%), these must be included because one cannot determine whether earlier surgery would have reduced this number. The challenge for the future is to identify predictors of success (or failure) before the onset of lytic infusion. It is generally accepted that successful lysis is more common when a guide-wire can be passed through the lesion into the distal circulation, and when
JOURNAL OF VASCULAR SURGERY February 1994
thrombosis is relatively fresh. In our data lytic success was not related to location of thrombosis or duration of occlusion; however, complication rates were significantly higher in patients with occlusions of more than 7 days duration. This difference persisted even when amputations were removed from the analysis. The reasons for this remain unclear and this observation must be confirmed or reported by other investigators. Although lysis is still possible in these cases, the riskjbenefit ratio seems to be increased in old occlusions. One of the potential advantages of thrombolytic therapy is to limit the extent of subsequent major surgical treatment. Our data did not support this hope in most patients. In the 33 cases in which no additional procedures were performed (group I), the long-term patency rate was similar to that reported by others in this setting4 and was significantly worse than that of the other treatment groups. This is not surprising because the underlying cause of thrombosis was not identified or treated. Therefore these patients cannot be considered to have long-term benefit from thrombolytic infusion. Patients in group I were routinely treated with warfarin (Coumadin) anticoagulation yet they still had very poor long-term patency rates. Nine cases were treated radiologically. Although data at 24 months follow-up are not available for this group of patients in our study, the IS-month patency rate was not different from that of the other treatment groups. The small number of cases treated solely by angioplasty reflects our selective use of that technique and the generally diffuse nature of the underlying disease. The results in this small group are encouraging. Surgical revascularization was simplified by lytic infusion in 22 patients (group III); although this was encouraging, these patients are only 16.4% of all patients treated. These patients had long-term patency rates similar to those with more extensive surgical procedures. However, the ability to perform angioplasty or minor surgical revision after lytic therapy did not confer improved long-term patency when compared with patients who required extensive surgical revascularization. Our data are consistent with the experience of others that showed only modest long-term patency rates after successful lysis. 2 -6 The extent of disease-causing graft thrombosis may be underestimated and is not often amenable to simple correction. Clearly in the case of a bypass graft it is better to identify the conduit in jeopardy before thrombosis. Analyzing the data in another way, by aggressive application of a lytic protocol we were able to avoid
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
extensive operation in 64 ofl34 cases (47.8%) with a 3.7% mortality rate, 20% complication rate but achieve no significant improvement in long-term patency. If the 14 cases in group I that thrombosed by 18 months are counted as failures, then the potential success rate drops to 50 of 134 (37.3%). Admittedly this is the harshest possible analysis of the data, and clearly some patients benefitted from the thrombolytic infusion by avoiding extensive surgical intervention. It appears from our data, however, that these patients are in the distinct minority. The cost-effectiveness of this therapy has been questioned before. 25 The challenge for future investigations is to identify characteristics that would predict success with minimal surgical intervention (groups II and III) or alternatively identify at the onset patients in whom extensive surgical revascularization seems to be required (groups IV and V). Unfortunately our data are too few to allow this identification to be made with certainty. Our analyses suggest that patients with native arterial thrombosis or with occlusion of suprainguinal arteries or grafts are more likely to be treated successfully by minimal intervention after lysis than those with graft or infrainguinal occlusion. Regrettably this is exactly the opposite of what one might hope to be the case, because these are the most easily and effectively treated surgically. Our experience suggests that thrombolytic therapy should not be applied routinely as first-line therapy for all peripheral thrombosis. Its expense, complication rate, and frequent need for extensive surgical reconstruction mitigate against its routine use. Assertions that major reconstructions in the face of acute ischemia are associated with increased short-term complications and reduced long-term patency or survival rates compared with PTA or minor revisions are not supported by our data. The major benefit of thrombolysis seems to be for the minority of patients in whom major reconstruction can be avoided. Future efforts should be directed at refining eligibility criteria for thrombolytic infusion by identifying those patients who could be treated by thrombolysis and minor intervention (groups II and III). Patients with acute native vessel occlusion limited to a single anatomic area or occlusion of one limb of a suprainguinal graft seem particularly suited for thrombolysis, although this remains to be proven. Selection of appropriate patients with infrainguinal graft occlusion for thrombolytic infusion remains problematic. Further prospective studies directed at specifics of conduit type and age, distal anatomy and
Faggioti et at.
295
extent of occlusive disease will be required to address this point. REFERENCES 1. Graor RA, Risius B, Young JR, et al. Thrombolysis of peripheral arterial bypass grafts: surgical thrombectomy compared with thrombolysis - A preliminary report. J VAse SURG 1988;7:347-55. 2. Belkin M, Donaldson MC, Whittemore AD, Polak: JF, Grassi CJ, Harrington DP, Mannick JA. Observations on the use of thrombolytic agents for thrombolytic occlusion of infrainguinal vein grafts. J VAse SURG 1990;11:289-96. 3. McNamara TO, Bomberger RA, Merchant RF. Intra-arterial urokinase as the initial therapy for acutely ischemic lower limbs. Circulation 1991;83[suppl 1]:1-106-119. 4. Gardiner GA, Harrington DP, Koltun W, Whittemore A, Mannick JA, Levin DC. Salvage of occluded by pass by means of thrombolysis. J VAse SURG 1989;9:426-3l. 5. van Breda A, Robison JC, Feldman L, et al. Local thrombolysis in the treatment of arterial graft occlusions. J VAse SURG 1984;1:103-12. 6. McNamara TO, Bomberger RA. Factors affecting initial and 6-month patency rates after intraarterial thrombolysis with high dose urokinase. Am J Surg 1986;152:709-12. 7. Wolfson RH, Kumpe DA, Rutherford RB. Role of intraarterial streptokinase in treatment of arterial thromboembolism. Arch Surg 1984;119:697-70l. 8. Sullivan K, Gardiner GA, Kandarpa K, et al. Efficacy of thrombolysis in infrainguinal bypass grafts. Circulation 1991; 83[SuppII:cb:I-99-105. 9. Ad Hoc Committee on Reporting Standards, SVSj NAISCVS. Rutherford RB, Preston Flanigan D, Gupta SK, et al. Suggested standards for reports dealing with lower extremity ischemia. J VAse SURG 1986;4:80-94. 10. McNamara TO, Fischer JR. Thrombolysis of peripheral graft occlusions: improved results using high-dose urokinase. Am J Roentgenol 1985;144:769-75. 11. Italian Cooperative Study Bologna. Endoarterial treatment of acute ischemia of the limbs with urokinase. Int Angiol 1989;8:53-6. 12. Sullivan KL, Gardiner GA, Shapiro MJ, Bonn J, Levin DC. Acceleration of thrombolysis with a high-dose transthrombus bolus technique. Radiology 1989;173:805-8. 13. Hallett ]W, Greenwood LH, Yrizarry JM, Pierson WP, Robison JG, Brown SB. Statistical determinants of success and complications of thrombolytic therapy for arterial occlusion of lower extremity. Surg Gynecol Obstet 1985;161: 431-7. 14. Faggioli GL, Ricotta JJ. Basic data: thrombolytic therapy for lower extremity occlusion. Ann Vasc Surg 1993;7:297-3l. 15. Belkin M, Belkin B, Buckman CA, Straub J], Lowe R. Intra-arterial fibrinolytic therapy: efficacy of streptokinase vs urokinase. Arch Surg 1986;121:769-73. 16. Cragg AH, Smith TP, Corson JD, et al. Two urokinase dose regimens in native arterial and graft occlusions: initial results of a prospective, randomized clinical trial. Radiology 1991; 178:681-6. 17. Ferguson LJ, Faris I, Robertson A, Lloyd JV, Miller JH. Intraarterial streptokinase therapy to relieve acute limb ischemia. J VAse SURG 1986;4:205-10. 18. McNamara TO. Role of thrombolysis in peripheral arterial occlusion. Am J Med 1987;83(suppI2A):6-10. 19. Hargrove WC, Berkowitz HD, Freiman DB, McLean G,
296 Faggioli et al.
20.
21. 22.
23. 24.
Ring EJ, Roberts B. Recanalization of totally occluded femoropopliteal vein grafts with low-dose streptokinase infusion. Surgery 1982;92:890-5. Browse DJ, Torrie EPH, Galland RB. Low dose intra-arterial thrombolysis in the treatment of occluded vascular grafts. Br J Surg 1992;79:86-8. Sicard GA, Schier JJ, Torry WG, et al. Thrombolytic therapy for acute arterial occlusion. J VAse SURG 1985;2:65-78. Ricotta JJ, Green RM, DeWeese JA. Use and limitations of thrombolytic therapy in the treatment of peripheral arterial ischemia: result of a multiinstitutional questionnaire. J VAse SURG 1987;6:45-50. Ricotta JJ. Intra-arterial thrombolysis. A surgical view. Circulation 1991;83[suppl 1]:1-120-121. Porter JM, Taylor LM. Current status of thrombolytic therapy. J VAse SURG 1985;2:239-49.
DISCUSSION Dr. Anthony D. Whittemore (Boston, Mass.). Since its inception, thrombolytic therapy has demonstrated potential as a short-term solution in search of the most appropriate clinical application. Regrettably, with respect to peripheral arterial occlusion, the literature consists of multicenter studies seriously flawed by several interrelated and confounding variables. Patient populations are heterogeneous; they include both arterial and native arterial graft occlusions. Both autogenous and prosthetic grafts are frequently studied together, along with both proximal and distal reconstructions. The duration of occlusion, which is always difficult to pinpoint with precision, varies from hours to months. Multiple agents are often combined and with varied dosages, including low, high, higher, and probably highest protocols, both with and without concurrent heparinization. Delivery systems vary widely from hand injections to powered and pulse-spray techniques, again with wide variation in how the initial bolus is administered. And finally, the lesions that are ultimately unmasked with thrombolytic therapy are handled variably, ranging from intervention with balloon angioplasty to formal reconstruction. The initial success rate of 65%, which is lower than that frequently published, is analogous to our own in every way with occluded vein grafts. The poor patency rates are also disappointingly similar. Do you believe that the outcome might be improved by limiting lytic therapy to the occlusions of very short duration, with more consistent and up-to-date doses? Can you elaborate on your recommendations for the patient population that is most likely to benefit from this? I am disturbed to see that three deaths were caused by bleeding into the brain. Would you elaborate on the conditions of those three patients to enable us to be more judicious in selecting our patients? Although proximal iliac occlusion seemed to be most
JOURNAL OF VASCULAR SURGERY February 1994
25. Dacey LJ, Dow RW, McDaniel MD, Walsh DB, Zwolak RM, Cronenwetr JL. Cost-effectiveness of intra-arterial thrombolytic therapy. Arch Surg 1988;123: 1218-23. 26. Perler BA, White RI, Ernst CB, Williams GM. Low-dose thrombolytic therapy for infrainguinal graft occlusions: an idea whose time has passed? J VAse SURG 1985;2:799805. 27. Seeger JM, Flynn TC, Quintessenza JA. Intra-arterial streptokinase in the treatment of acute arterial thrombosis. Surg GynecolObst 1987;164:303-7. 28. Rush DS, Gewerts BL, Lu CT, et al. Selective infusion of streptokinase for arterial thrombosis. Surgery 1983;93:82833.
Submitted June 10, 1993; accepted Aug. 10, 1993.
favorable, many of us, including myself, believe that the most expeditious way to handle this complication is through a limited groin incision, with the patient receiving local anesthetic in the operating room and thereby avoiding all the preoperative intensive care unit monitoring and intracerebral disease and death. What is your recommendation for the way to handle an acute iliac occlusion? Would you elaborate on the theoretic advantage with regard to dissolution of thrombus in the distal tibial peroneal outflow tract? Were you able to analyze for that with regard to its influence on limb salvage? Thus we agree with your overall conclusion that lysis should not be routinely used in a knee-jerk fashion for peripheral arterial occlusion, and we concur with your plea that continued and well-controlled studies are desperately needed. Dr. Kenneth Ouriel (Rochester, N.Y.). The vascular community is gradually recognizing that the goal of thrombolysis is not to replace or eliminate surgery but rather to diminish the magnitude of required interventions and thereby decrease the morbidity and mortality rates. In this regard, Dr. Ricotta's group observed satisfactory long-term patency rates with thrombolysis plus adjuvant operation, patency rates that do not differ significantly from those of operation alone. Although the absence of a suitably matched surgical group renders the justification of the authors' conclusion questionable, our data seem to corroborate their opinion. Patency, however, is not the only parameter with which to gauge success. Long-term morbidity and mortality rates are important, and the Buffalo group study is admirably different from previous studies on the basis of long-term follow-up. Even the recently completed GUSTO trial of streptokinase and tissue-type plasminogen activator in more than 30,000 patients with acute myocardial infarction fell short in this regard, basing conclusions on early,
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
in-hospital outcome. The early mortality rate in the Buffalo study was a little more than 3% in patients treated with thrombolytic therapy. This is strikingly lower than corresponding operative mortality rates of more than 25% in Blaisdell's review, 15% to 20% in Jivegaard's, and more than 10% in the recently reported series from Oregon. What was the observed long-term survival rate in the total group of 134 patients treated with thrombolysis? And, if it was substantially lower than the reported rates after operation alone, do these results not justify the use of thrombolysis on the basis of identical patency rates coinciding with improved patient survival rates? Dr. John J. Ricotta. Regarding which patients in whom this will be useful, I believe that thrombolysis will be useful in patients who have single-segment or limited occlusions on their angiograms. I believe we can identify these patients as patients who, after lysis, will be able to undergo PTA. An iliac artery lesion limited to the common iliac artery, for example, would be an artery on which I would try lysis. When thrombosis occurs in a graft with a known or suspected focal stenosis, or in a graft with early vein graft failure, when the vein was believed to be adequate by inspection at the time of the operation or completion angiography, you might expect successful lysis there. I am concerned about late vein graft failures because you may find a more diffuse problem that is not likely to allow you to perform a revision.
Paggioti et at.
297
In a vein graft wherein there is a normal angiography result, if that vein fails, there is usually a diffuse ischemic insult when the graft fails. I doubt that thrombolysis would be helpful in those patients. The other group has those patients in whom restoring the conduit patency allows you to perform a more limited procedure, preferably one where you could use saphenous vein. Therefore if someone had an above-knee femoropopliteal vessel that occluded, and it was possible to use a piece of vein to extend it to a lower popliteal segment, it might be worthwhile. But I believe we need to develop these criteria. We need to examine these patients, not only to decide what operation we would perform if we did not have thrombolysis, but what operation we will perform after the thrombolytic therapy is successful. We could not predict death in these patients, and this has been a real concern. These patients did not have prolonged infusions. The people that died had infusions less than 18 hours. With regard to the distal thrombus, I recommend Dr. Comerota's data that suggest that intraoperative lytic therapy can lyse a lot of thrombus even in patients with chronic occlusive disease. We did not find significant late mortality rates in our patients. It is interesting to speculate that perhaps systemic lysis is connected with reduced late mortality rates.
Intrarterial thrombolytic therapy with urokinase is commonly used for the treatment of acute and subacute thrombosis of arteries and grafts. The rationale for this approach is based on good early results, acceptable complication rates, the possibility of minimizing subsequent interventions, and the theoretical capacity to dissolve clots in the smaller peripheral vessels. l . S There are few data in the From the Department of Surgery, Division of Vascular Surgery, State University of New York at Buffalo, and Department of Vascular Surgery (Drs. Pedrini, Di Paola, and D'Addato), University of Bologna. Presented at rhe Forty-sevenrh Annual Meeting of rhe Society for Vascular Surgery, Washington, D.C., June 8-9, 1993. Reprint requests: John J. Ricotta, MD, Department of Surgery, Division of Vascular Surgery, Millard Fillmore Hospital, 3 Gates Circle, Buffalo, NY 14209. Copyright © 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, Norrh American Chapter. 0741-S214(94($3.00 + 0 24/6/50763
literature that address the long-term fate of vessels successfully recanalized by thrombolytic infusion, with or without the use of adjunctive procedures. l ,2,4.8 Without these data the long-term benefit of thrombolysis as a therapeutic intervention will be difficult to assess. To obtain some of this information, we examined the long-term results obtained in a population of patients treated with intraarterial urokinase therapy for lower limb arterial occlusions.
MATERIAL AND METHODS The records of all thrombolytic infusions performed for lower limb arterial or graft thrombosis over a 7-year period were reviewed. All patients were diagnosed with class I and II ischemia with Society for Vascular Surgery/International Society for Cardiovascular Surgery, North American Chapter criterion. 9 Urokinase was used as thrombolytic agent and 289
JOURNAL OF VASCULAR SURGERY February 1994
290 Faggioli et at.
Table 1. Early results of thrombolytic therapy in arteries and grafts No.
Arteries Suprainguinal Infrainguinal All Grafts Suprainguinal Infrainguinal All Total Suprainguinal Infrainguinal All
Failures
Complication
Death
4 (21.0%) 15 (38.5%) 19 (32.7%)
4 (21%) 10 (25.6%) 14 (24.1%)
1 (2.5%)
2 (8%) 11 (21.5%) 13 (17.1%)
1 (1.9%)
9 (11.9%)
7 (28%) 21 (41.2%) 28 (36.8%)
5 (11.4%) (12.3%) 16 (11.9%)
11 (25%) 36 (40%) 47 (35.1%)
6 (13.6%) 21 (23.3%) 27 (20%)
3 (6.8%) 2 (2.2%) 5' (3.7%)
CL
PL
19 39 58
12 (63.2%) 20 (51.3%) 32 (55.2%)
± (10.2%)
25 51 76
16 (64%) 23 (45.1%) 39 (51.3%)
Z (13.8%)
44 90 134
28 (63.6%) 43 (47.7%) 71 (52.9%)
3 (15.8%)
7 (12.1%)
2 (8%)
11
0(0%)
1 (1.7%) 3 (12%)
4 (5.2%)
P > 0.05 between rows.
delivered intraarterially with the tip of the catheter inserted into the thrombus whenever possible. Over the course of time, different infusion protocols were followed. These include the protocol described by McNamara and FischerlO (initial dose of 4000 IV/min followed by a maintenance dose of 1000 IV/min), and other protocols described in the literature (75,000 to 150,000 IV bolus dose with 25,000 to 75,000 IV/hr maintenance dose).)),12 Concomitant intravenous heparin infusion was used in most cases. Angiograms were obtained before, during, and at completion of the thrombolytic treatment. Results of the lytic infusions were classified as complete lysis (CL, restoration of flow throughout the vessel with no evidence of residual clot), partial lysis (PL, achievement of patency with minimal residual clot inside the lumen), and failure (absence of or minimal clot lysis). The patient population was then divided in five groups according to the result of the lytic treatment and the type of subsequent treatment performed for analysis of long-term results: group I, success of thrombolytic therapy (CL or PL) with no evidence of correctable lesions and no additional therapy; group II, success of thrombolytic therapy with minor defects in the lumen corrected by percutaneous trans luminal angioplasty (PTA); group III, success of thrombolytic therapy with a localized residual defect (a limited surgical procedure [i.e., angioplasty, jump graft] was performed); group IV, success of thrombolytic therapy with extensive residual disease of the vessel or the graft (an extensive surgical procedure [i.e., new bypass, extensive proximal or distal reconstruction] was required in these cases]; group V, failure of lytic therapy and surgical revascularization. Follow-up was obtained by direct examination and vascular laboratory evaluation of the patient. Immediate patency rates were compared by chi-
squared analysis. Differences between late results were analyzed by life-table analysis (Kaplan-Meier) and compared by log-rank test (Mantel-Haenszel). Statistical significance was defined as p < 0.05.
RESULTS Complete information on 134 urokinase infusions in 122 patients was available. There were 76 (56.7%) grafts (25 suprainguinal and 51 infrainguinal) and 58 (43.3%) native arteries (19 suprainguinal and 39 infrainguinal). The mean duration of symptoms was 4.6 ± 8.8 days (median 1.5), and the mean infusion time was 53.8 hours ± 29.0 (median 48). Early results Complete lysis was obtained in 71 cases (52.9%) and partial lysis in 16 (11.9% ), for a total of 8 7 of 134 (64.9%) successful cases. A complete description of early results in grafts and arteries according to their location is reported in Table I. No statistical difference in efficacy or mortality and morbidity rates was found by chi-squared analysis between graft and native artery or between infrainguinal and suprainguinallocation (p > 0.05) (Table I). Five deaths occurred for an overall mortality rate of 3.7%. The causes of death were directly related to the infusion therapy and included cerebral hemorrhage (three cases), myocardial infarction (one case), and gastrointestinal bleeding (one case). Major complications occurred in 27 cases (20.1%): these included 11 amputations as a result offailed thrombolytic therapy and prolonged ischemia, 10 hematomas that required surgical evacuation or blood transfusion, two peripheral embolizations, one case of pulmonary edema, one myocardial infarction, one lacunar infarction, and one catheter thrombosis. Mortality and morbidity rates were evenly distributed
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
Paggioli et al.
291
Table II. Early complications and long-term patency (± SE) in treatment groups Complications*
Group I
Perioperative
6 mas
12 mas
18 mas
24 mas
25.90/0 ± 21.0%
1 (3%)
96.9% ± 3.3%
51.0% ± 9.5%
39.3% ± 12.4%
25.9% ± 12.5%
0
77.8% ± 12.3%
77.8% ± 18.5%
58.3% ± 26.7%
58.3% ± 37.6%t
3 (13.6%)
81.8% ± 7.5%
63.0% ± 11.5%
63.0% ± 14.4%
63.0% ± 19.1%t
63.0% ± 27.0%t
3 (13.0%)
91.3% ± 5.7%
78.2% ± 10.1%
56.3% ± 13.1%
56.3% ± 16.6%t
56.3% ± 21.4%t
3 (14.3%)
80.0% ± 7.8%
57.2% ± 11.8%:j:
51.8% ± 13.6%:j:
41.4% ± 14.1%:j:
41.4% ± 18.1%:j:
(n = 33)
Group II (n = 9)
Group III (n = 22) Group IV (n = 23) Group V (n = 21)
*Group I, myocardial infarction; group III, one embolization and two hematomas; group IV, one embolization and two hematomas; and group V, three hematomas. tp < 0.05 compared with group 1. :j:p = NS compared with other groups.
in the different graft types and anatomic categories. Previous investigators have correlated duration of infusion (> 12 to 24 hours) with systemic lytic effect. 13 In our patients the duration of treatment, that is, less or more than 18 hours, did not significantly influence incidence of death or disease. The time from the occlusive event to the beginning of lytic treatment was recorded in 101 cases. Fourteen patients (13.8%) were treated after more than 7 days of occlusion, with CL in six (42.8%) patients, PL in two (14.3%), and failures in six (42.8%) patients. In the 87 patients (86.1%) treated within a week from occlusion there was CL in 38 (55.2%) patients, PL in 12 (13.8%), and failures in 27 (31%) patients. The difference between these groups (8 ofl4 = 57.1%; 50 of87 = 57.5%) is not statistically significant (p > 0.25). However, the morbidity and mortality rates in the group with longer duration of occlusion were significantly greater (71.4% = five complications, four amputations, one death) than the rates of the group with an occlusion time less than 7 days (17.2% = five complications, SIX amputations, four deaths) (p < 0.001). Long-term results The distribution of cases by result of lytic therapy was as follows. There were 33 cases (24.6%) in group I, in which no additional procedures were performed because of the inability to detect a correctable lesion. Twenty-seven (81.9%) of these cases showed complete lysis and six (18.1 %) had some minor residual thrombi. All patients in this group underwent anticoagulant therapy. The presence of residual clot in these six cases did not seem to influence subsequent follow-up, because only one case with PL occluded within 1 month and the others remained patent for at
least 6 months. The residual stenosis after lytic therapy was treated by PTA in nine cases (6.7%, group II). Twenty-two cases (16.4%) were treated by "minor" surgical procedures, including 10 jump grafts, 10 surgical angioplasties, and two thromboembolectomies (group III). In 23 other cases (17.2%) an "extensive" surgical procedure was performed (group IV): in seven cases this was an aortobifemoral bypass and in 16 cases a new femoropopliteal or femorodistal bypass. Twenty-one cases ( 15 .7% ) in which the lytic therapy failed were treated surgically (group V); the intervention performed was a femoropopliteal or femorodistal bypass in 10 cases, a thromboembolectomy in five cases, a thromboembolectomy with angioplasty in three cases, and an axillofemoral bypass in three cases. There was no difference in perioperative complications between any of the treatment groups (Table II). Twenty-six cases (19.4%) are not included in the five groups listed above. They consist of the five deaths and the 11 primary amputations that occurred during lytic treatment and 10 other patients who had clinical signs of improvement in the ischemic limb despite unsuccessful thrombolytic treatment. Three of these patients were lost to follow-up, three were admitted again with rest pain at 7, 18, and 36 months; and four patients had limiting claudication at 5, 6, 16, and 18 months. Overall, thrombolysis was effective in avoiding major vascular reconstruction in 64 of 134 cases (47.8%). Follow-up was available in 74 of 108 (68.6%) cases included in the five treatment groups, for a mean of 10.9 ± 10.8 months (range 1 to 60 months). Life-table analyses of each group of treatment are reported in Table II and in Fig. 1). Results in group I were significantly inferior to those of the other groups of intervention after successful throm-
JOURNAL OF VASCULAR SURGERY February 1994
292 Faggioli et al.
100
80
-- GROUP I
.., > U Z
I
-0-
.•.•.-t ............................... ..
60
L4-...j.QooQ-ooQool~i-'---..g.... - - -
w
GROUP II
.... GROUP III
- -.
~
-- GROUP IV 40
-- GROUP V
20
23 5 14 17 15
19 5 13 16 14
14 4 11 13 10
9 4 11 11 10
8
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
0 0
2 11
6 2 7
6 1 4
3 1 4
8
8
8
8
7
7
7
5
4 5 5
2 3 3
2 3 3
2 3 3
2
2
GI Gil Gill GIV GV
FOLLOW-UP (Months) Fig. 1. Life-table analyses in five treatment groups (group I, success with no further therapy; group II, percutaneous angioplasty alone; group III, limited surgical procedure; group IV, extensive procedure; and group V, revascularization after lytic failure).
bolysis (groups II, III, and IV). This was true both at the 18-month interval (group I vs group II, p = 0.01; group I vs group III or IV,p < 0.05) and at the 24-month interval (group I vs group III or IV, p < 0.05). When surgical intervention was required, long-term patency was similar regardless of the extent of the procedure used (group III vs group IV, p < 0.25 at 18 and 24 months). Patients who underwent surgical reconstruction after failure of thrombolysis (group V) demonstrated a reduced long-term patency (41% at 24 months), although differences with the other treatment groups were not statistically significant at 18 and 24 months. Patency rates in patients who underwent "minor" treatment after thrombolysis (groups II and III = 62.9% at 18 and 24 months) were not significantly different from those of patients in whom an "extensive" intervention was required, either after successful or unsuccessful thrombolysis (groups IV and V = 53.2% at 18 and 24 months; p > 0.25) (Fig. 2). Regardless of the treatment performed after the lytic therapy, grafts and arteries demonstrated similar long-term patency rates (58.7% ± 12.5% vs 56.6% ± 18.6% at 18 months,p > 0.25). In native arteries, cases in groups II and III ("minor" inter-
vention) had better long-term results than those in groups IV and V ("extensive" treatment) (100% vs 55.8% at 18 months), although this approached but did not reach statistical significance (p = 0.06). Native arteries also showed better patency rates than grafts in groups II and III (100% vs 55.8%, P = 0.06) (Table III). Suprainguinal grafts tended to behave better than infrainguinal grafts after a 12-month interval (p = 0.06), but this effect was lost at 18 and 24 months (p = 0.1). No significant difference was found when 18- and 24-month patency rates of suprainguinal arteries were compared with patency rates of infrainguinal arteries (p > 0.1). However, the long-term patency rates in groups II and III was significantly greater in suprainguinal (91.6% at 18 months) than in infrainguinal grafts (39.6%; p < 0.025). In suprainguinal grafts groups II and III also had better 18-month patency rates than groups IV and V (32.7%;p < 0.05) (Table III). Saphenous veins had the lowest 6- and 12-month patency rates (38.3% ± 11.3% and 25.5% ± 15.5%), but these results were not significantly different from that of polytetrafluoroethylene grafts (64.1 % ± 15.7% and 51.3% ± 20.6%;p > 0.1).
JOURNAL OF VASCULAR SURGERY Volwne 19, Number 2
Faggioli et al.
293
100
80
-- GROUP I
:>
u 60
GROUP II & III
Z
-0-
~
.... GROUP IV & V
w
Q.
40
20
0
0
23 19 32
19 14 18 15 30 23
2
4
6
9 8 6 15 13 9 21 15 15
6
5 15
1 3 5 4 13 10
1 2
1 2
1 2
6
6
6
2
2
GI GII&III GIV&V
8 10 12 14 16 18 20 22 24 26 28 30
FOLLOW-UP (Months) Fig. 2. Comparison oflife-table analyses in group I (success with no further therapy); groups II and III ("minor" procedure after successful thrombolysis); and groups IV and V ("extensive" procedure after successful or failed thrombolysis).
DISCUSSION Intraarterial, preferably intrathrombus, urokinase infusion has been proposed as the first-choice treatment for acute arterial and graft thrombosis. Immediate success rates with intrathrombus urokinase can be as high as 88% in both arteries and grafts,3-8 and a metaanalysis of the English literature reveals an average success rate of85% in the most recent series, with a complication rate of 20%.14 There are many potential attractions of thrombolytic therapy over an initial surgical approach. They include the ability to restore patency with a percutaneous procedure, the possibility of improved patency of run-off vessels with the increased ability to plan subsequent intervention, and the potential to simplify subsequent surgical procedures. 1-5,15-18 Some authors state that the very distal arterial bed can be cleared of small thrombi more efficiently than with balloon embolectomy, without jeopardizing the endothelial integrity.l,2 In aggregate, these potential advantages have been proposed to justify thrombolysis as the procedure of choice and are believed by many to offset its complication rate and cost. 4 ,5,15-19 Acceptance of thrombolysis as the procedure of choice has not been universal. Skeptics point out the
Table III. Long-term results in groups II and III ("minor" procedures) and groups IV and V ("extensive" procedures)
Infrainguinal Suprainguinal Arteries Grafts <7 days ocelusion >7 days ocelusion
*p
=
Groups II andlli
Groups IV and V
p Value
39.6%* 91.6%* 100%* 55.8%* 58.0%
62.4 32.7% 45.8% 60.0% 48.1%
NS <0.05 <0.05 NS NS
60.0%
30.0%
NS
0.06 within treatment groups.
cost and complications of the technique and frequent need for adjunctive surgical procedures. 2o-28 No prospective randomized data are available. In a series of 27 cases treated with streptokinase, Wolfson, Kumpe, and Rutherford7 found an I8-month patency rate of 22% and 50% in grafts and arteries, respectively. Graor et al. 1 retrospectively compared 33 patients who underwent thrombolysis with 38 cases treated surgically. Thirty-day patency rate was significantly higher in the lytic treatment group (86%
294 Faggioli et al.
vs 42%), but the I-year patency rates were 20% and 14%, a difference that is unlikely to be significant in such a small population. 1 McNamara and Bomberger6 found a 6-month patency rate of 59% ina group of 100 cases treated with high-dose urokinase; patency rates of vein grafts after successful thrombolysis were 36.6% and 22.9% at 1 and 2 years in the study from Belkin et aI.2 The most favorable results were reported by Sullivan et al.,8 and were as high as 79% patency at 2 years in the cases with a correctable lesion after successful lytic treatment. If thrombolysis cannot be shown to improve long-term vessel patency and limb salvage, its routine use with attendant risks and expense will be difficult to justify. Therefore an evaluation oflong-term patency after lytic infusion is critical. There must be clear benefit associated with thrombolytic infusion to offset its cost and complication rate. Infusion must either reduce short-term morbidity rates, mortality rates, and cost without compromising long-term patency or provide improved long-term patency without significant increase in short-term complications. Our study was designed to retrospectively review the short-term complications and long-term patency associated with the use of thrombolytic infusion. Whereas analysis of these data is limited by the constraints of a retrospective study, it can provide insight about conditions in which thrombolysis is of potential benefit. Our practice has been to use thrombolytic infusion in all the patients with acute thrombotic occlusion, in the absence of specific contraindications. Despite different protocols for urokinase delivery, we could find no relationship of infusion protocols to complication rate or results in our series. The retrospective nature of our review and variety of doses used precludes any conclusion in this regard. We did not see increased morbidity rates associated with operative intervention. Although our success rate was somewhat lower than that reported by some authors, it is comparable to other experiences in the literature. 2 ,16,25 Thrombolysis was successful in 65% of cases, associated with a 3.7% mortality rate and a 20% complication rate. Although about one third of the complications were amputations (S.2%), these must be included because one cannot determine whether earlier surgery would have reduced this number. The challenge for the future is to identify predictors of success (or failure) before the onset of lytic infusion. It is generally accepted that successful lysis is more common when a guide-wire can be passed through the lesion into the distal circulation, and when
JOURNAL OF VASCULAR SURGERY February 1994
thrombosis is relatively fresh. In our data lytic success was not related to location of thrombosis or duration of occlusion; however, complication rates were significantly higher in patients with occlusions of more than 7 days duration. This difference persisted even when amputations were removed from the analysis. The reasons for this remain unclear and this observation must be confirmed or reported by other investigators. Although lysis is still possible in these cases, the riskjbenefit ratio seems to be increased in old occlusions. One of the potential advantages of thrombolytic therapy is to limit the extent of subsequent major surgical treatment. Our data did not support this hope in most patients. In the 33 cases in which no additional procedures were performed (group I), the long-term patency rate was similar to that reported by others in this setting4 and was significantly worse than that of the other treatment groups. This is not surprising because the underlying cause of thrombosis was not identified or treated. Therefore these patients cannot be considered to have long-term benefit from thrombolytic infusion. Patients in group I were routinely treated with warfarin (Coumadin) anticoagulation yet they still had very poor long-term patency rates. Nine cases were treated radiologically. Although data at 24 months follow-up are not available for this group of patients in our study, the IS-month patency rate was not different from that of the other treatment groups. The small number of cases treated solely by angioplasty reflects our selective use of that technique and the generally diffuse nature of the underlying disease. The results in this small group are encouraging. Surgical revascularization was simplified by lytic infusion in 22 patients (group III); although this was encouraging, these patients are only 16.4% of all patients treated. These patients had long-term patency rates similar to those with more extensive surgical procedures. However, the ability to perform angioplasty or minor surgical revision after lytic therapy did not confer improved long-term patency when compared with patients who required extensive surgical revascularization. Our data are consistent with the experience of others that showed only modest long-term patency rates after successful lysis. 2 -6 The extent of disease-causing graft thrombosis may be underestimated and is not often amenable to simple correction. Clearly in the case of a bypass graft it is better to identify the conduit in jeopardy before thrombosis. Analyzing the data in another way, by aggressive application of a lytic protocol we were able to avoid
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
extensive operation in 64 ofl34 cases (47.8%) with a 3.7% mortality rate, 20% complication rate but achieve no significant improvement in long-term patency. If the 14 cases in group I that thrombosed by 18 months are counted as failures, then the potential success rate drops to 50 of 134 (37.3%). Admittedly this is the harshest possible analysis of the data, and clearly some patients benefitted from the thrombolytic infusion by avoiding extensive surgical intervention. It appears from our data, however, that these patients are in the distinct minority. The cost-effectiveness of this therapy has been questioned before. 25 The challenge for future investigations is to identify characteristics that would predict success with minimal surgical intervention (groups II and III) or alternatively identify at the onset patients in whom extensive surgical revascularization seems to be required (groups IV and V). Unfortunately our data are too few to allow this identification to be made with certainty. Our analyses suggest that patients with native arterial thrombosis or with occlusion of suprainguinal arteries or grafts are more likely to be treated successfully by minimal intervention after lysis than those with graft or infrainguinal occlusion. Regrettably this is exactly the opposite of what one might hope to be the case, because these are the most easily and effectively treated surgically. Our experience suggests that thrombolytic therapy should not be applied routinely as first-line therapy for all peripheral thrombosis. Its expense, complication rate, and frequent need for extensive surgical reconstruction mitigate against its routine use. Assertions that major reconstructions in the face of acute ischemia are associated with increased short-term complications and reduced long-term patency or survival rates compared with PTA or minor revisions are not supported by our data. The major benefit of thrombolysis seems to be for the minority of patients in whom major reconstruction can be avoided. Future efforts should be directed at refining eligibility criteria for thrombolytic infusion by identifying those patients who could be treated by thrombolysis and minor intervention (groups II and III). Patients with acute native vessel occlusion limited to a single anatomic area or occlusion of one limb of a suprainguinal graft seem particularly suited for thrombolysis, although this remains to be proven. Selection of appropriate patients with infrainguinal graft occlusion for thrombolytic infusion remains problematic. Further prospective studies directed at specifics of conduit type and age, distal anatomy and
Faggioti et at.
295
extent of occlusive disease will be required to address this point. REFERENCES 1. Graor RA, Risius B, Young JR, et al. Thrombolysis of peripheral arterial bypass grafts: surgical thrombectomy compared with thrombolysis - A preliminary report. J VAse SURG 1988;7:347-55. 2. Belkin M, Donaldson MC, Whittemore AD, Polak: JF, Grassi CJ, Harrington DP, Mannick JA. Observations on the use of thrombolytic agents for thrombolytic occlusion of infrainguinal vein grafts. J VAse SURG 1990;11:289-96. 3. McNamara TO, Bomberger RA, Merchant RF. Intra-arterial urokinase as the initial therapy for acutely ischemic lower limbs. Circulation 1991;83[suppl 1]:1-106-119. 4. Gardiner GA, Harrington DP, Koltun W, Whittemore A, Mannick JA, Levin DC. Salvage of occluded by pass by means of thrombolysis. J VAse SURG 1989;9:426-3l. 5. van Breda A, Robison JC, Feldman L, et al. Local thrombolysis in the treatment of arterial graft occlusions. J VAse SURG 1984;1:103-12. 6. McNamara TO, Bomberger RA. Factors affecting initial and 6-month patency rates after intraarterial thrombolysis with high dose urokinase. Am J Surg 1986;152:709-12. 7. Wolfson RH, Kumpe DA, Rutherford RB. Role of intraarterial streptokinase in treatment of arterial thromboembolism. Arch Surg 1984;119:697-70l. 8. Sullivan K, Gardiner GA, Kandarpa K, et al. Efficacy of thrombolysis in infrainguinal bypass grafts. Circulation 1991; 83[SuppII:cb:I-99-105. 9. Ad Hoc Committee on Reporting Standards, SVSj NAISCVS. Rutherford RB, Preston Flanigan D, Gupta SK, et al. Suggested standards for reports dealing with lower extremity ischemia. J VAse SURG 1986;4:80-94. 10. McNamara TO, Fischer JR. Thrombolysis of peripheral graft occlusions: improved results using high-dose urokinase. Am J Roentgenol 1985;144:769-75. 11. Italian Cooperative Study Bologna. Endoarterial treatment of acute ischemia of the limbs with urokinase. Int Angiol 1989;8:53-6. 12. Sullivan KL, Gardiner GA, Shapiro MJ, Bonn J, Levin DC. Acceleration of thrombolysis with a high-dose transthrombus bolus technique. Radiology 1989;173:805-8. 13. Hallett ]W, Greenwood LH, Yrizarry JM, Pierson WP, Robison JG, Brown SB. Statistical determinants of success and complications of thrombolytic therapy for arterial occlusion of lower extremity. Surg Gynecol Obstet 1985;161: 431-7. 14. Faggioli GL, Ricotta JJ. Basic data: thrombolytic therapy for lower extremity occlusion. Ann Vasc Surg 1993;7:297-3l. 15. Belkin M, Belkin B, Buckman CA, Straub J], Lowe R. Intra-arterial fibrinolytic therapy: efficacy of streptokinase vs urokinase. Arch Surg 1986;121:769-73. 16. Cragg AH, Smith TP, Corson JD, et al. Two urokinase dose regimens in native arterial and graft occlusions: initial results of a prospective, randomized clinical trial. Radiology 1991; 178:681-6. 17. Ferguson LJ, Faris I, Robertson A, Lloyd JV, Miller JH. Intraarterial streptokinase therapy to relieve acute limb ischemia. J VAse SURG 1986;4:205-10. 18. McNamara TO. Role of thrombolysis in peripheral arterial occlusion. Am J Med 1987;83(suppI2A):6-10. 19. Hargrove WC, Berkowitz HD, Freiman DB, McLean G,
296 Faggioli et al.
20.
21. 22.
23. 24.
Ring EJ, Roberts B. Recanalization of totally occluded femoropopliteal vein grafts with low-dose streptokinase infusion. Surgery 1982;92:890-5. Browse DJ, Torrie EPH, Galland RB. Low dose intra-arterial thrombolysis in the treatment of occluded vascular grafts. Br J Surg 1992;79:86-8. Sicard GA, Schier JJ, Torry WG, et al. Thrombolytic therapy for acute arterial occlusion. J VAse SURG 1985;2:65-78. Ricotta JJ, Green RM, DeWeese JA. Use and limitations of thrombolytic therapy in the treatment of peripheral arterial ischemia: result of a multiinstitutional questionnaire. J VAse SURG 1987;6:45-50. Ricotta JJ. Intra-arterial thrombolysis. A surgical view. Circulation 1991;83[suppl 1]:1-120-121. Porter JM, Taylor LM. Current status of thrombolytic therapy. J VAse SURG 1985;2:239-49.
DISCUSSION Dr. Anthony D. Whittemore (Boston, Mass.). Since its inception, thrombolytic therapy has demonstrated potential as a short-term solution in search of the most appropriate clinical application. Regrettably, with respect to peripheral arterial occlusion, the literature consists of multicenter studies seriously flawed by several interrelated and confounding variables. Patient populations are heterogeneous; they include both arterial and native arterial graft occlusions. Both autogenous and prosthetic grafts are frequently studied together, along with both proximal and distal reconstructions. The duration of occlusion, which is always difficult to pinpoint with precision, varies from hours to months. Multiple agents are often combined and with varied dosages, including low, high, higher, and probably highest protocols, both with and without concurrent heparinization. Delivery systems vary widely from hand injections to powered and pulse-spray techniques, again with wide variation in how the initial bolus is administered. And finally, the lesions that are ultimately unmasked with thrombolytic therapy are handled variably, ranging from intervention with balloon angioplasty to formal reconstruction. The initial success rate of 65%, which is lower than that frequently published, is analogous to our own in every way with occluded vein grafts. The poor patency rates are also disappointingly similar. Do you believe that the outcome might be improved by limiting lytic therapy to the occlusions of very short duration, with more consistent and up-to-date doses? Can you elaborate on your recommendations for the patient population that is most likely to benefit from this? I am disturbed to see that three deaths were caused by bleeding into the brain. Would you elaborate on the conditions of those three patients to enable us to be more judicious in selecting our patients? Although proximal iliac occlusion seemed to be most
JOURNAL OF VASCULAR SURGERY February 1994
25. Dacey LJ, Dow RW, McDaniel MD, Walsh DB, Zwolak RM, Cronenwetr JL. Cost-effectiveness of intra-arterial thrombolytic therapy. Arch Surg 1988;123: 1218-23. 26. Perler BA, White RI, Ernst CB, Williams GM. Low-dose thrombolytic therapy for infrainguinal graft occlusions: an idea whose time has passed? J VAse SURG 1985;2:799805. 27. Seeger JM, Flynn TC, Quintessenza JA. Intra-arterial streptokinase in the treatment of acute arterial thrombosis. Surg GynecolObst 1987;164:303-7. 28. Rush DS, Gewerts BL, Lu CT, et al. Selective infusion of streptokinase for arterial thrombosis. Surgery 1983;93:82833.
Submitted June 10, 1993; accepted Aug. 10, 1993.
favorable, many of us, including myself, believe that the most expeditious way to handle this complication is through a limited groin incision, with the patient receiving local anesthetic in the operating room and thereby avoiding all the preoperative intensive care unit monitoring and intracerebral disease and death. What is your recommendation for the way to handle an acute iliac occlusion? Would you elaborate on the theoretic advantage with regard to dissolution of thrombus in the distal tibial peroneal outflow tract? Were you able to analyze for that with regard to its influence on limb salvage? Thus we agree with your overall conclusion that lysis should not be routinely used in a knee-jerk fashion for peripheral arterial occlusion, and we concur with your plea that continued and well-controlled studies are desperately needed. Dr. Kenneth Ouriel (Rochester, N.Y.). The vascular community is gradually recognizing that the goal of thrombolysis is not to replace or eliminate surgery but rather to diminish the magnitude of required interventions and thereby decrease the morbidity and mortality rates. In this regard, Dr. Ricotta's group observed satisfactory long-term patency rates with thrombolysis plus adjuvant operation, patency rates that do not differ significantly from those of operation alone. Although the absence of a suitably matched surgical group renders the justification of the authors' conclusion questionable, our data seem to corroborate their opinion. Patency, however, is not the only parameter with which to gauge success. Long-term morbidity and mortality rates are important, and the Buffalo group study is admirably different from previous studies on the basis of long-term follow-up. Even the recently completed GUSTO trial of streptokinase and tissue-type plasminogen activator in more than 30,000 patients with acute myocardial infarction fell short in this regard, basing conclusions on early,
JOURNAL OF VASCULAR SURGERY Volume 19, Number 2
in-hospital outcome. The early mortality rate in the Buffalo study was a little more than 3% in patients treated with thrombolytic therapy. This is strikingly lower than corresponding operative mortality rates of more than 25% in Blaisdell's review, 15% to 20% in Jivegaard's, and more than 10% in the recently reported series from Oregon. What was the observed long-term survival rate in the total group of 134 patients treated with thrombolysis? And, if it was substantially lower than the reported rates after operation alone, do these results not justify the use of thrombolysis on the basis of identical patency rates coinciding with improved patient survival rates? Dr. John J. Ricotta. Regarding which patients in whom this will be useful, I believe that thrombolysis will be useful in patients who have single-segment or limited occlusions on their angiograms. I believe we can identify these patients as patients who, after lysis, will be able to undergo PTA. An iliac artery lesion limited to the common iliac artery, for example, would be an artery on which I would try lysis. When thrombosis occurs in a graft with a known or suspected focal stenosis, or in a graft with early vein graft failure, when the vein was believed to be adequate by inspection at the time of the operation or completion angiography, you might expect successful lysis there. I am concerned about late vein graft failures because you may find a more diffuse problem that is not likely to allow you to perform a revision.
Paggioti et at.
297
In a vein graft wherein there is a normal angiography result, if that vein fails, there is usually a diffuse ischemic insult when the graft fails. I doubt that thrombolysis would be helpful in those patients. The other group has those patients in whom restoring the conduit patency allows you to perform a more limited procedure, preferably one where you could use saphenous vein. Therefore if someone had an above-knee femoropopliteal vessel that occluded, and it was possible to use a piece of vein to extend it to a lower popliteal segment, it might be worthwhile. But I believe we need to develop these criteria. We need to examine these patients, not only to decide what operation we would perform if we did not have thrombolysis, but what operation we will perform after the thrombolytic therapy is successful. We could not predict death in these patients, and this has been a real concern. These patients did not have prolonged infusions. The people that died had infusions less than 18 hours. With regard to the distal thrombus, I recommend Dr. Comerota's data that suggest that intraoperative lytic therapy can lyse a lot of thrombus even in patients with chronic occlusive disease. We did not find significant late mortality rates in our patients. It is interesting to speculate that perhaps systemic lysis is connected with reduced late mortality rates.