A strategy of aggressive regional therapy for acute iliofemoral venous thrombosis with contemporary venous thrombectomy or catheter-directed thrombolysis

A strategy of aggressive regional therapy for acute iliofemoral venous thrombosis with contemporary venous thrombectomy or catheter-directed thrombolysis Anthony J. Comerota, MD, Samuel C. Aldridge, MD, Gary Cohen, MD, David S. Ball, DO, Mark Pliskin, MD, and John V. White, MD, Philadelphia and East Stroudsburg, Pa. Purpose: Occlusive iliofemoral venous thrombosis is associated with morbid short- and long-term consequences. Having been disappointed with standard anticoagulant therapy and systemic fibrinolysis, we embarked on an aggressive multidisciplinary regional approach to treat these patients, with the goals of therapy being (1) to eliminate iliofemoral venous thrombus, (2) to provide unobstructed venous drainage from the affected limb, and (3) to prevent recurrent thrombosis. Methods: Twelve consecutive patients were treated for extensive iliofemoral venous thrombosis. Each had thrombus from their infrapopliteal veins through their ifiofemoral system, and four had vena caval involvement. The conditions of 11 patients failed to improve when the patients were given anticoagulants, and prior systemic fibrinolysis failed in five patients. The treatment strategy includes catheter-directed thrombolysis with intrathrombus infusion of the plasminogen activator or operative thrombectomy or venous bypass with a permanent 4 mm arteriovenous fistula (AVF). Results: Nine of 12 patients had a good or excellent clinical outcome (mean follow-up 25 months), which correlated with restored unobstructed venous drainage from the affected limb. Seven patients had catheter-directed lyric therapy attempted. In five patients the catheters were appropriately positioned, and lysis was successful. Five of the eight patients who underwent operations had successful procedures. Two of the three patients with poor operative outcomes had residual thrombus in their iliac veins or vena cava after thrombectomy (without bypass). The third patient, in whom anticoagulation was contraindicated, had an initially successful thrombectomy and AVF; however, vena caval thrombosis developed 2 months after operation. N o patient had symptomatic pulmonary emboli, and routine posttreatment ventilation/perfusion lung scanning was not performed. Conclusions: An aggressive multidisciplinary regional approach to patients with obliterarive iliofemoral venous thrombosisi designed to remove thrombus and provide unobstructed venous drainage, offers substantially better clinical outcome compared with systemic fibrinolysis and standard anticoagulation. Catheter-directed thrombolysis is successful if the catheter is appropriately positioned within the thrombus. Contemporary venous thrombectomy, which includes thrombus removal, completion phlebography, AVF, and cross-pubic bypass when necessary, is associated with high success rates. Failures can be anticipated and avoided in most patients. (J VASC SURG 1994;20:244-54.) From the Sectionof Vascular Surgery(Drs. Comerota, Aldridge, and White), and the Section of Interventional Radiology (Dr. Cohen and Mr. Ball),TempleUniversityHospital,Philadelphia, and the Department of Radiology (Dr. Pliskin), Pocono Medical Center, East Stroudsburg. Supported in part by NIH grant No. 1 K07 HL02658-01. Presented at the Fifth Annual Meeting of the AmericanVenous Forum, Orlando, Fla., Feb. 24-26, 1993. Reprint requests: Anthony J. Comerota, MD, Department of Surgery, Broad and Ontario St., Philadelphia, PA 19140. Copyright © 1994 by The Society for Vascular Surgery and International Societyfor CardiovascularSurgery,North American Chapter. 0741-5214/94/$3.00 + 0 24/6/55746 244

Patients with iliofemoral deep venous thrombosis (DVT), many o f whom are categorized as having phlegmasia cerulea dolens, have a high risk o f acute pulmonary emboli (PE), have significant acute morbidity, and are likely to have severe postthrombotic sequelae. 1,2 Physicians observing the morbid natural history ofiliofemoral D V T have an intuitive desire to eliminate the thrombus, restoring venous drainage from the leg. Historically, this was initially attempted with venous thrombectomy. Although preliminary reports were encouraging, 3,4 long-term follow-up o f

JOURNAL OF VASCULAR SURGERY Volume 20, Number 2

Comerota et al.

245

I ILIOFEMORAL DVT I ICONTRAINDICATIONS TO LYTlC THERAPY [

VENOUS THROMBECTOMY~ WITH AVF/BYPASS

|

I

ArrEMPT CATHETER DIRECTED THROMBOLYSIS

]

I' "c T. .APY i Fig. 1. Algori~m for treating iliofemoral venous thrombosis.

patients who underwent operation failed to demonstrate objective benefit, s,6 Operative morbidity also became an issue, with patients having PE, blood loss often requiring transfusion, prolonged hospitalization, and an appreciable operative mortality rate. 5 As a result, venous thrombectomy has not been accepted as reasonable treatment for these patients, despite more recent reports indicating substantially better outcome. 7,8 Thrombolyric therapy is an alternative that has the potential for pharmacologically restoring patency to the affected veins. Although systemic lyric therapy for acute DVT is not widely accepted, when used it is reserved for patients with phlegmasia cerulea dolens. We adopted a similar policy in patients with phlegmasia certflea dolens who had no contraindication to lyric therapy; however, the results of systemic thrombolysis were marginal in some patients and poor in most. Because we observed extreme postthrombotic morbidity with iliofemoral DVT despite aggressive anticoagulation, and because our experience with systemic lyric therapy had been disappointing, we adopted an aggressive regional approach for the treatment of these patients. The goal of this approach is to provide unobstructed venous return to the affected limb, primarily by restoring patency to the iliofemoral venous system by taking advantage of the more efficient delivery of thrombolyric agents via

regional (intrathrombus) infusion and the improved techniques of contemporary venous thrombectomy. Previous reports regarding treatment ofiliofemoral venous thrombosis review singular therapy. During the past 6 years we adopted a treatment strategy for patients with occlusive iliofemoral venous thrombosis, with use of catheter-directed intrathrombus infusion offibrinolytic agents, venous thrombectomy with arteriovenous fistula (AVF), selective venous bypass, or a combination of these techniques. With this multimodality approach, eliminating the severe acute morbidity and avoiding long-term postthrornboric sequelae can be anticipated, because treatment options are expanded and concepts are unified. The purpose of this study is to present the aggressive, regional therapy of patients with occlusive iliofemoral venous thrombosis and to review our initial results in 12 patients treated in this manner. MATERIAL AND M E T H O D S Between May 1988 and May 1993, 12 patients were admitted with occlusive iliofemoral venous thrombosis. Ages ranged from 20 to 58 years, with a mean of 40 years. The average duration of leg symptoms was 4.4 days (range 2 to 10 days), with an average of 4.1 days (2 to 8 days) for patients who underwent operation and 4.6 days (2 to 10 days) for those treated with catheter-directed thrombolysis. All

246

Comerota et al.

IOURNAL OF VASCULAR SURGERY August 1994

Fig. 2. Example of catheter-directed thrombolysis for acute iliofemoral venous thrombosis extending to popliteal vein in 40-year-old man presenting 9 days after total colectomy for ulcerative colitis. A, Iliocavagram (anteroposterior and oblique views) demonstrates thrombus extending from left iliac vein partially occluding distal vena cava. B, Bird's Nest vena caval filter was placed above thrombus, and two multi-side-hole catheters were inserted into clot, one from right jugular vein and one from contralateral femoral vein. Bolus rt-PA and continuous infusion of urokinase was given through both catheters to restore venous patency. C, After 72 hours, patency was restored with excellent clinical response. Note embolus trapped by caval filter (arrow). At 1 year superficial femoral veins and iliofemoral veins remained patent. patients had the diagnosis objectively confirmed with venous duplex imaging and iliofemoral phlebography. Venous duplex imaging was used to evaluate the infrainguinal deep venous system. Iliofemoral phlebography was performed to assess the proximal extent of thrombus, and routine contralateral iliocavagraphy evaluated the contralateral iliofemoral system and vena cava. Although three patients were known to have recent PE, the remainder had no symptoms of PE. Although seven patients underwent

pretreatment ventilation/perfusion scanning, routine posttreatment ventilation/perfusion scanning was not performed. I f a patient had a PE during the acute thrombotic episode, a vena caval filter was inserted. If the patient had large, irregular, and nonocclusive vena caval clots, avena caval filter was also inserted. We have not routinely placed caval filters before operation or before catheter-directed thrombolysis if the thrombus was limited to the iliofemoral venous system. However, six of our patients had caval filters

JOURNALOF VASCULARSURGERY Volume 20, Number 2

Comerota et al. 247

Fig. 3. Technique of venous thrombectomy and AVF. A, Preoperative ascending phlebogram of 38-year-old woman with development of early phlegmasia cerulea dolens 6 days after spine reconstruction for scoliosis. All named veins, from her foot to vena cava are thrombosed. B, Through longitudinal femoral incision, common femoral, saphenous, and superficial femoral veins are exposed. Transverse venotomy is made in common femoral vein, (arrow), Which is packed with thrombns (Left). Within short time, thrombus begins to extrude from venotomy, (double arrow) because of high venous pressure (Right). C, Leg is raised and rubber bandage is wrapped tightly from foot to upper thigh, to remove as much clot as possible from infrainguinal venous system (Top). After passage of no. 10 venous thrombectomy catheter proximally, one can appreciate extensive amount of thrombus retrieved (Bottom). D, Completion venogram demonstrates patent iliofemoral venous system without residual thrombus or obstruction (Left). Small (4 mm) AVF (arrow) is created, sewing end of transsected saphenous vein (or large proximal branch) to side of superficial femoral artery (Right). Proximal saphenous vein frequently requires thrombectomy before creating A-V fistula. Small cuff of 5 mm PTFE graft around proximal saphenous vein segment is now used routinely. E, Photograph taken at 3-year follow-up visit. Patient has only mild intermittent swelling, controlled with low-pressure gradient compression stockings.

248

Comerota et al.

IOURNAL OF VASCULAR SURGERY August 1994

Fig. 4. This panent is example of combination therapy with catheter-directed intraarterial thrombolysis, venous thrombectomy, and cross-pubic venous bypass. A, Ascending phlebography in patient with advanced phlegmasia cerulea dolens and impending venous gangrene demonstrates thrombosis of every named vein of his right lower extremity. Left iliofemoral venous system and vena cava were patent and free of thrombus. Arteriogram was obtained to evaluate arterial inflow because of impending gangrene. There was no occlusive disease, however, because of severe venous hypertension, no contrast was visualized distal to popliteal artery despite delayed imaging. Catheter was left in place, and urokinase was infused into common femoral artery at 4000 U/min. B, Graph of patient's venous pressure recorded via dorsal foot vein, indicates severe venous hypertension initially. Drop in pressure was rapid in response to urokinase infusion, which plateaued at 78 cm 1-I20. Although patient's pain resolved, swelling persisted. After 16 hours without additional improvement, urokinase was discontinued and patient was taken to operating room for venous thrombectomy. No thrombus was extracted from veins below inguinal ligament. Although thrombus was removed from iliofemoral venous system, persistent pelvic venous obstruction was appreciated, therefore cross-pubic bypass was performed. Patient's venous pressure returned to normal after operation and patient was given anticoagulants. C, Predischarge phlebogram demonstrates patent deep venous system with multiple functional valves in superficial femoral vein and patent cross-pubic bypass (arrows). Although there was some degree of residual venous obstruction of popliteal veto, patient's noninvasive physiologic study results were normal, without evidence of venous insut~ciency.

in place at the time o f treatment. Three patients had vena caval thrombosis Occluding above a previously placed filter, and three patients had a caval filter inserted as part o f current therapy. Patients who had no contraindication to thrombolytic therapy had an attempt at catheter-directed thrombolysis. Those in whom the catheter-directed approach failed and those with an absolute contraindication to thrombolytic therapy or multiple relative contraindications underwent a venous thrombectomy and AVF receiving a general anesthetic (Fig. 1). A cross-pubic venous bypass was constructed if tmobstructed ipsilateral iliofemoral venous patency could not be restored. The cause for acute venous thrombosis was established in all patients. Postoperative/posttraumatic thrombosis occurred in seven patients, malig-

nancy was fotmd in four, and an underlying hypercoagulable state was fotmd in one patient. Patient follow-up was 8 to 63 months (mean 25 months). One patient was lost to follow-up after 36 months, and two patients died because o f metastatic cancer 2 weeks and 8 months after therapy. Catheter-directed thrombolysis Direct intraclot infusion is achieved by placing a catheter from the contralateral femoral vein, the right jugular vein, or both (Fig. 2). An alternative access through an ipsilateral femoral vein catheter can be used at the discretion o f the physician. Use o f the contralateral femoral vein or the jugular vein allows placement o f a vena caval filter, if indicated, before infusion. Although we have not routinely inserted vena caval filters before catheter-directed lyric

JOURNAL OF VASCULAR SURGERY Volume 20, Number 2

Comerota et al.

249

Table I. Patient characteristics, treatment, and outcome Treatment

Patient~Side Age~Sex 1. R i g h t 2. Left 3. Left 4. 5. 6. 7, 8, 9, 10, 11 12.

Left Right Left Right Bilat. Bilat. Left Right Left

Cause

Anticoagulant

20/F 27/M 39/F

Post-op Trauma Malignancy

+ -

34/F 57/M 36/M 36/M 54/M 58/M 40/M 54/F 21/M

Post-op Mafignancy Trauma Post-op Hypercoag Malignancy Post-op Malignancy Trauma

1+ -

Unobstructed venous Follow-up Clinical* Thrombolysis Thrombectomy Bypass A VF drainage (months) outcome Failed sys Intra-op C a t h dir (Failed sys) None C a t h dir Failed sys$ C a t h dir C a t h dir Failed sys Cath dir None None

~-

-

~ + -

Yes Yes Yes

23 40 8

Excellent Good Good ±

+ +

+ § -

+ ~-

Yes Yes No Yes Yes No Yes Yes Yes

54 63 40 19 20 0.5 12 10 8

Excellent Excellent Poor Excellent Excellent Poor ± Excellent Good,q?oor Good

Post-op, P o s t o p e r a t i v e ; Failed O,s, failed systemic: Intra-op, intraoperative; Cath dir, catheter-directed; Hypercoag, hypercoagulable. * O u t c o m e evaluation: Excellent = m i n i m a l o r n o s y m p t o m s , external c o m p r e s s i o n n o t required, g r a d e 0; Good - s w e l l i n g c o n t r o l l e d w i t h g r a d i e n t c o m p r e s s i o n stockings, g r a d e 1; Poor = v e n o u s claudication, u n c o n t r o l l e d s w e l l i n g p o s t o p e r a t i v e or persistent pain, grade 2. TPatient died o f m e t a s t a t i c carcinoma. ± A t t e m p t s at i n t r a t h r o m b u s catheter p l a c e m e n t failed. § R e q u i r e d bypass 2 m o n t h s after t h r o m b e c t o m y because o f d e v e l o p m e n t o f iliac vein stenosis (appeared to be n e o i n t i m a l fibroplasia). A l t h o u g h initial results w e r e g o o d , anticoagulants w e r e c o n t r a i n d i c a t e d and vena caval t h r o m b o s i s d e v e l o p e d in the patient.

therapy, they were inserted in three patients; two had pulmonary embolism as part of their current presentation, and one had large, nonocclusive caval thrombus. The indications for caval filtration in this setting are currently evolving. Patients requiring thrombectomy subsequent to caval filtration present technical limitations that can compromise operative results. Once a guide wire is appropriately positioned in the thrombus, multi-side-hole catheters are advanced into the thrombus to assure maximal delivery of the lyric agent to the fibrin-bound plasminogen. Systemic doses of plasminogen activators are used. We have chosen to use urokinase delivered as a 250,000to 500,000-unit bolus followed by continuous infusion of 250,000 U/hr. In two patients bolus doses of recombinant tissue plasminogen activator (rt-PA) (10 mg) followed by urokinase infusion (250,000 U/hr) were used to take advantage of potential synergistic fibrinolysis. 911 Venous duplex imaging is used to follow lysis of the infrainguinal clot and as much of the iliac venous system as can be visualized. Repeat phlebography through the infusion catheters is performed at 12- to 24-hour intervals, and therapy is continued until maximal lysis is achieved. Routine coagulation studies are performed before treatment. During treatment prothrombin times, partial thromboplastin times, fibrinogen, and fibrin(ogen) split products are monitored at 12-hour intervals. If the fibrinogen level falls below 100 mg/dl, the lyric agent is stopped for 6 to 12 hours, and additional heparin is given. Lyric infusion is resumed when the fibrino-

gen level rises above 100 mg/dl. After completion of the lyric infusion, patients continue to receive heparin and, later, oral anticoagulants. If the infusion catheters cannot be appropriately positioned within the iliac vein thrombus, a venous thrombectomy is performed (Fig. 3). Venous thrombectomy Although a venous thrombectomy can be performed with the patient receiving a local anesthetic, regional or general anesthesia is preferred. Because most patients are receiving heparin and continue to receive anticoagulants after operation, regional (spinal or epidural) anesthesia is contraindicated. Our preference is general endotracheal anesthesia with at least 10 cm of positive end-expiratory pressure applied. Patients have two units of blood cross-. matched and autotransfusion devices are used routinely to minimize transfusion requirements. A vertical inguinal incision is made over the femoral vessels and the common, superficial femoral, deep femoral, and saphenous veins are mobilized and controlled. Patients are given systemic anticoagulant with heparin at 100 U/kg (or more). A transverse or slightly oblique venotomy is made in the distal common femoral vein just proximal to the saphenofemoral junction (Fig. 3). The proximal thrombus is removed with passage of a no. 8 to 10 venous thrombectomy catheter. A balloon catheter is not passed from the contralateral femoral vein to occlude the vena cava before thrombectomy because the

JOURNAL OF VASCULAR SURGERY August 1994

250 Comerota et al.

Table II, A. Results of aggressive regional approach for iliofemoral DVT: catheter-directed thrombolysis Succes6rul catheter positioning

Successful lysis Yes No Clinical outcome Good to excellent Poor

Yes

No

52 0

0 2

5~ 0

0 2t

Intrafemoral artery infusion in one patient for impending venous gangrene, and intraaortic and intracaval infusion in another for acute aortic thrombosis complicating phlegmasia cerulea dolens. tTwo patients with failed infrainguinal clot lysis underwent operative thrombectomy, and one patient with successful infrainguinal clot lysis had a venous bypass performed because of persistant iliac vein obstruction.

incidence ofpulrnonary embolism is low. 12,13Positive end-expiratory pressure is applied during the thrombectomy, further minimizing the likelihood of PE. In an awake patient the Valsalva maneuver can be performed. If the patient has a vena caval filter in place, fluoroscopy is used to guide the passage of the thrombectomy catheter, and liquid contrast is used to inflate the balloon. The proximal venous segment is always assessed by completion phlebography, obtained with an injection of 25 to 30 ml of contrast material through the common femoral venotomy. In light of the inflow occlusion, this technique generally gives good visualization of the entire iliac venous system (Fig. 3, C). An alternative is to use direct venoscopy with an angioscope, as reported by Loeprecht. 14We have not been successful with venous angioscopy because of our inability to adequately clear the lumen of blood draining from collateral veins. If patency cannot be restored to the proximal venous segment, or if it is found to be extrinsically compressed, a cross-pubic venous bypass is performed with an 8 to 10 mm externally supported polytetrafluoroethylene (PTFE) graft. A 3 to 4 mm AVF is created approximately 4 to 6 cm distal to the origin of the superficial femoral artery (Fig. 4). After the proximal thrombectomy, the distal thrombus is removed. As much clot as possible is extracted by the extrusion technique, either by stroking along the course of the distal veins or by use of a tight Esmarch bandage wrapped from the foot proximally. Attempts can be made to pass a Fogarty venous thrombectomy catheter distally, although its passage is usually prohibited by venous valves. Others

have reported exploration of the posterior tibial vein into which a catheter is introduced and advanced to the common femoral vein. A thrombectomy catheter can then be attached and guided distally, at which time a thrombectomy of the entire leg can be performed. 12 Although we have no experience with this technique, it may represent an improvement over what can be achieved by compression/extrusion alone; however, the endothelial damage created by balloon catheters is always a concern. Venography can be performed through the distal point of access to assess the adequacy of thrombectomy. An AVF is constructed with a large proximal branch of the greater saphenous vein, or the proximal greater saphenous itself if a branch is not available. The proximal greater saphenotts vein usually requires thrombectomy to restore patency; therefore one does not "sacrifice" it for the AVF. If only the proximal segment is thrombosed, every effort is made to preserve the vein by a proximal thrombectomy and creating the AVF to a branch, if available. A small piece of 4 mm PTFE graft is placed around the vein before anastomosis with the superficial femoral artery. This ensures that the arteriovenous communication will not enlarge to create a vascular steal. The superficial femoral artery is mobilized 4 to 6 cm distal to its origin, and an end-to-side anastomosis, no greater than 4 mm in diameter, is performed. A double loop of prolene suture is passed around the vein segment (PTFE graft) with a silver or titanium clip joining its ends and left in the subcutaneous tissue. This can be readily accessed in the future, should one wish to close the AVF. Because the AVFs are only 3 to 4 mm in diameter, they do not cause hemodynamic consequences, and therefore routine closure is not required. Heparin is continued throughout the operative and postoperative period, and warfarin (Coumadin) therapy is begun on the first postoperative day. External pneumatic compression devices are applied in the recovery room to further accelerate venous return and prevent recurrent thrombosis. Clinical outcome was defined as excellent if patients had no postthrombotic signs or symptoms, did not require gradient compression stockings to control swelling, and were classified as grade 0 in the Society for Vascular Surgery/International Society for Cardiovascular Surgery (SVS/ISCVS) classification of chronic venous disease) s A good result was defined as mild postthrombotic symptoms with gradient compression stockings required to control swelling and classified as SVS/ISCVS grade 1. Apoor result is defined as persistent pain, poorly controlled

JOURNAL OF VASCULARSURGERY Volume 20, Number 2

Comerota et al.

251

Table II, B. Results of aggressive regional approach for iliofemoral DVT: operative therapy Clinical outcome

Operative outcome

Good to excellent (6 patients)

Successful thrombectomy, AVF, bypass if residual iliac vein destruction Unsuccessful thrombectomy, no AVF, no bypass

Poor (2 patients)

swelling despite gradient compression stockings, symptoms of venous claudication, and SVS/ISCVS grade 2 or 3.

RESULTS The results of treatment were categorized according to clinical outcome and the SVS/ISCVS clinical classification of chronic venous disease. Outcome correlated directly with patency of the iliofemoral venous segment or the venous bypass. Twelve patients with occlusive iliofemoral venous thrombosis were treated (Table I). Each had thrombus extending from the infrapopliteal veins through the iliac systems, and four had vena caval involvement. The conditions of 11 patients did not improve with anticoagulation, and one patient had anticoagulation withheld because of an intracranial malignancy. Previous systemic fibrinolysis failed in five patients; two separate courses of intravenous thrombolysis failed in one patient. Iliofemoral thrombus (three involving the vena cava) was successfully treated in nine of 12 patients, with patency restored or a patent bypass providing unobstructed venous outflow from the affected limb. These patients had either a good or excellent clinical outcome. Two of the treatment failures were persistent occlusion of the iliofemoral veins and vena cava, and one venal caval thrombosis that developed 2 months after operation. Seven patients had catheter-directed lyric therapy attempted (Table II). In five of the seven patients, the catheters were appropriately positioned, and in each of these five successful lysis occurred, resulting in a good or excellent clinical outcome. One patient underwent intraarterial thrombolysis with urokinase via the femoral artery for unilateral iliofemoral DVT with impending venous gangrene. In a second patient a complex problem of massive phlegmasia cerulea dolens complicated by acute aortic thrombosis developed. Multiple catheters were positioned; two were positioned intraarterially (distal aorta and left iliac artery) and one intravenous catheter was embedded into the intracaval thrombus. This patient was treated with bolus doses of rt-PA, followed by a continuous infusion of urokinase through both the arterial and venous catheters. The two patients

having intraarterial infusion had excellent and good clinical outcomes, respectively. Although both patients who had unsuccessful catheter placement were treated with urokinase, fibrinolysis failed in both, and the patients underwent venous thrombectomy, which was successful in one and unsuccessful in the other.

Eight patients had venous thrombectomy, five as a primary procedure, and three after catheter-directed lysis. Four had an AVF and three underwent venous bypass to provide unobstructed venous drainage from the affected limb. Catheter-directed lysis failed in two of the eight patients who underwent operation, and one patient underwent an adjunctive iliofemoral thrombectomy and bypass after successful lysis of infrainguinal thrombus. Two patients had residual obstruction after iliofemoral thrombectomy, one in the proximal common iliac vein and one with residual iliocaval thrombus. Another patient with an intracranial malignancy initially had a good outcome from thrombectomy with AVF but required crosspubic bypass within 3 weeks because of iliofemoral stenosis believed to be intimal fibroplasia. She did well for 3 months when accelerated leg swelling subsequently developed because ofvena caval thrombosis (anticoagulation was contraindicated). The bypass and AVF were dismantled. These three patients were classified in the poor clinical outcome group. In all five patients in whom unobstructed venous outflow was restored, a good or excellent result was obtained (mean follow-up 38 months). No patient had symptomatic pulmonary embolism dur-ing therapy or hospitalization. The cross-pubic venous bypass of one patient was dismantled because of vena caval thrombosis; two patients with cross-pubic venous bypass were monitored with long-term venous duplex scanning for 40 and 63 months. Both had patent bypasses and remained symptom free. Venous function studies (air plethysmography) were obtained during follow-up on all patients capable of performing the study. Two patients died of metastatic cancer, and two were unable to perform the test. Of the remaining eight patients, three were hemodynamically normal at 23, 40, and 63 months,

252 Comerom et al.

Fig. 5. Schematic of preferred method of cross-pubic venous bypasswith 8 or 10 mm externallysupported PTFE graft. Note small AVF (-<4 ram) to superficial femoral artery.

four patients had mild to moderate venous insufficiency, and two had severe insufficiency. DISCUSSION

Traditional teaching indicates that patients with phlegmasia cerulea dolens are adequately treated with bedrest, anticoagulation, and leg elevation. However, personal observation indicates that this is not the case, with most patients having substantial acute morbidity and severe postthrombotic sequela. The severe morbidity of iliofemoral venous thrombosis has been long recognized?,2 Furthermore, studies evaluating venous function after thrombosis indicate that the degree of the post-thrombotic sequelae is related to the magnitude of the acute thrombosis, with these patients representing the most extreme examples of acute DVT. 16 The pathophysiologic condition of the postthrombotic syndrome is ambulatory venous hypertension, with valvular incompetence and luminal obstruction as its pathologic components. The most

JOURNAL OF VASCULAR SURGERY August 1994

severe manifestations of the postthrombotic syndrome associated with the highest ambulatory venous pressures occur in patients who have both obstruction and valvular incompetence.17 Recent studies demonstrate that valvular function can be preserved after physiologic lysis of DVT, especially when it occurs over a relatively short period) 8,19 Extending these observations to therapeutic fibrinolysis, one could expect further preservation of longterm valvular function with successful lysis. This has been demonstrated in three prospective randomized series. 2°22 For the most part, patients facing postthrombotic sequelae after thrombolytic therapy are those in whom lysis fails or who have recurrent thrombosis, not those in whom lysis is successful. Unfornmately, our experience with systemic thrombolysis for iliofemoral venous thrombosis has been disappointing, and others have had similar resultsY This is not surprising, because the iliofemoral venous system is "packed" with thrombus. When blood flow is obliterated, plasminogen activator can not reach the thrombus, and therefore lysis does not occur. The regional, intrathrombus delivery of thrombolytic agents takes advantage of the basic principle ofthrombolytic therapy, which is activation of fibrin-bound plasminogen. 24 All patients with appropriate intrathrombus catheter positioning had successful fibrinolysis and did not require additional intervention. One patient with intraarterial lytic infusion for thrombosis of every named vein in his leg and pelvis had successful clearing of his infrainguinal thrombus but required thrombectomy and venous bypass for persistent occlusion of his iliofemoral venous segment. No patient had a clinically evident PE during therapy or the remainder of their hospitalization. Although a vena caval filter was in place in six of our patients at the time of treatment, three were previously placed and were occluded with caval thrombosis at presentation. Although vena caval filters present a technical challenge during venous thrombectomy, catheter-directed thrombolysis around and through caval filters has not created difficulties. Although concerned about the potential for bleeding during lysis through sites of filter attachment to the cava, transmural bleeding was not observed in any of these patients. Molina and colleagues25 recently reported their experience with catheter-directed thrombolysis for iliofemoral venous thrombosis. They successfully treated 10 of 12 patients and found that 6 of the 10 who underwent lysis required balloon dilation or stent placement to treat a residual iliac vein stenosis.

JOURNAL OF VASCULAR SURGERY Volume 20, Number 2

We have not treated any of the patients in this series with balloon dilation or a venous stent; however, four patients could have been considered potential candidates if transluminal repair of their residual iliac vein lesion was considered. Two of our patients with particularly severe phlegmasia cerulea dolens, who had vena caval thrombosis complicating their iliofemoral DVT, were treated with a combination of rt-PA and urokinase. Ten-milligram pulsed infusions of rt-PA were given, followed by a continuous infusion of urokinase at 4000 U/min. Both patients had remarkable lysis, requiring no further intervention. The rationale for this combination is the synergistic lytic activity9-n caused by their different mechanisms of action and fibrin binding capability. Successful iliofemoral venous thrombectomy is now the rule rather than the exception. Rutherford 26 has reviewed the historical perspective of venous thrombectomy, and emphasized the evolution of improved results with the more recent series. 7,8 The principles of contemporary venous thrombectomy include complete proximal clot removal, removal of as much distal thrombus as possible, assessment of the iliofemoral venous system for residual thrombus or an underlying lesion (proximal iliac vein web or extrinsic compression), the creation of a small AVF or the construction of a cross-pubic venous bypass (also with a small AVF) to achieve adequate venous outflow. If patients are selected properly for thrombectomy (within 7 days), clearing of the iliofemoral venous system should not present a problem. Once patency is restored, measures must be taken to ensure that rethrombosis does not occur,. Completion iliofemoral phlebography is essential, allowing the surgeon to identify and correct any residual problem in the most appropriate fashion. Therapeutic anticoagulation coupled with the increased flow velocity of a small AVF has been shown to significantly improve results.7 Because recurrent thrombosis can occur after closure of the AVF, and because a small AVF does not cause hemodynamic problems, we do not close AVFs routinely. Only one patient required subsequent closure because of swelling and severe venous hypertension resulting from thrombosis of the vena cava. We also favor the application of external pneumatic compression garments after operation, to further reduce the chance of recurrent thrombosis. The risk of causing a PE as a result of external pneumatic compression is unsubstantiated, and its benefit appears to outweigh the risk. Intraluminal venous pressures with simple foot dorsi and plantar flexion are many times higher than that obtained with

Comerota et al.

253

a 60 to 80 mm pneumatic pressure garment externally applied. ~ Thrombectomy in three of eight patients was considered a failure. Two predictable failures were caused by inability to dear the venous system during the initial procedure, and in the third patient vena caval thrombosis developed 2 months after operation. Analyzing these data, it is evident that the key to a successful outcome is removal of the iliofemoral venous thrombus, restoration of unobstructed venous drainage from the involved leg, and prevention of recurrent thrombosis. Although our preference is to restore patency and preserve the native venous system, there are instances, such as chronic organized thrombus, an intraluminal web, or iliac vein compression, when this is not possible. If the contralateral iliofemoral system and the vena cava are patent, unilateral obstruction is easily handled with a cross-pubic venous bypass. We have found the saphenous vein to be inadequate because of its small size relative to the deep venous system, and in our experience the saphenous vein frequently failed when used for cross-pubic venous bypass. Previous reports of successful cross-pubic venous bypasses with saphenous veins focused on patients with chronic iliac vein o b s t r u c t i o n . 27 In many patients the pelvic vein obstruction was caused by malignancy, frequently presenting as a preterminal event without associated iliofemoral thrombosis. Patients with malignant pelvic venous obstruction are different than those described in this report, in that obstruction is gradual and a thrombectomy is usually not required. A recent report by G r u s s 28 indicates improved patency with a PTFE cross-pubic venous bypass compared with saphenous vein. Our preferred bypass conduit is an externally supported PTFE graft at least 8 mm in diameter. When constructed with an inline, contiguous 3 to 4 mm AVF (Fig. 5) combined with long-term anticoagulation, sustained patency can be anticipated. Balloon dilation of residual lilac vein stenosis (with stenting if appropriate) can also be considered if fluoroscopy is available. Despite the extensive degree of venous thrombosis in these patients, the acute relief of the iliofemoral venous occlusion eliminated early morbidity in 10 of the 12 patients (83%) and provided sustained benefit in 9 of 12 (75%). The patients reported in this study represent the most extensive acute venous thrombosis cases pre~Personal observation during lower extremity venous pressure measurements, 1979-1981.

254

Comerota et al.

senting during a 5-year period. Each had multilevel venous occlusion with thrombus extending from the lower leg through the common iliac vein, with 33% involving the vena cava. This approach is conceptually unified, and designed to eliminate thrombus, restore patency, provide unobstructed venous return and prevent rethrombosis. An aggressive regional, multidisciplinary approach offers to patients with the most severe form of DVT the possibility of substantially improved early and long-term prognosis compared with anticoagulation alone. REFERENCES 1. O'Donnell TF, Browse NL, Burnand KG, et all The socioeconomic effects of an iliofemoral thrombosis. J Surg Res 1'977;22:483-8. 2. Cockett FB, Thomas L. The iliac compression syndrome. Brit J Surg 1965;52:816-21. 3. Hailer JA, Abrams BL. Use ofthrombectomy in the treatment of acute ifiofemoral venous thrombosis in forty-five patients. Ann Surg 1963;158:561-6. 4. DeWeese J'A. Thrombectomy for acute iliofemoral venous thrombosis. J Cardiovasc Surg 1964;5:703-12. 5. Lansing AM, Davis WM. Five-year follow-up study of iliofemoral venous thrombectomy. Ann Surg 1968;168: 620-8. 6. Karp RB, Wylie EL Recurrent thrombosis after iliofemoral venous thrombectomy. Surg Forum 1966;17:147-8. 7. Plate G, Einarsson E, Ohlin P, Jensen R, Qvarfordt P, Eklof B. Thrombectomy with temporary arteriovenous fistula: the treatment of choice in acute iliofemoral venous thrombosis. J VASC SURG 1984;1:867-76. 8. Swedenborg J, Hagglof R, Jacobsson H, et aL Results of surgical treatment for iliofemoral venous thrombosis. Br J Surg 1986;73:871-4. 9. Gurewich V, Pannell R. A comparative study of the efficacy and specificity of tissue plasminogen activator and prourokinase: demonstration o f synergism and of different thresholds of non-selectivity. Thromb Res 1986;44:217-21. 10. Collen D, Stump D, Van de WerfF. Coronary thrombolysis in patients with myocardial infarction by intravenous infusion of synergic thrombolytic agents. Am Heart J 1986;11: 1083-7. 11. Collen D, Stassen J, Suamp D, Verstraete M. Synergism of thrombolytic agents in vivo. Circulation 1986;14:838-42. 12. Eklof B, Juhan C. Revival of thrombectomy in the management of acute iliofemoral venous thrombosis. Contemp Surg 1992;40:21-30.

JOURNAL OF VASCULARSURGERY August 1994

13. Kistner RL, Sparkuhl MD. Surgery in acute and chronic venous disease. Surgery 1979;85:31-43. 14. Loeprecht H. Angiosopie veineuse. Phlebologie 1988;41: 165-70. 15. Porter JM, Rutherford RB, Clagett GP, et al. Reporting standards in venous disease. J VAsc SURG 1988;8:172-81. 16. Lindner DJ, Edwards JM, Phinney ES, Taylor LM, Porter JM. Long-term hemodynamic and clinical sequelae of lower extremity deep vein thrombosis. J VAsc SURG 1986;4:43642. 17. Shull KC, Nicolaldes AN, Fernandes e Fernandes J, et al. Significance of popliteal reflux in relation to ambulatory venous pressure and ulceration. Arch Surg 1979; 114:1304-6. 18. Killewich LA, Bedford GR, Beach KW, Strandness DE. Spontaneous lysis of deep venous thrombi: rate and outcome. J VAsc SURG 1989;9:89-97. 19. Meissner MH, Manzo RA, Bergelln RO, Markel A, Strandness DE. Deep venous insufficiency: the relationship between lysis and subsequent reflux. J VAsc SURG 1993;18:596-608. 20. Elliot MS, Immelman EJ, Jeffrey P, et al. A comparative randomized trial of heparin versus streptokinase in the treatment of acute proximal venous thrombosis: an interim report of a prospective trial. Br J Surg 1979;66:838-42. 21. Arnesen H, Hoiseth A, Ly B. Streptokinase or heparin in the treatment of deep vein thrombosis: follow-up results of a prospective study. Acta Med Scand 1982;211:65-71. 22. Jeffrey P, Immelman E, Amoore J. Treatment of deep vein thrombosis with heparin or streptokinase: long-term venous function assessment. London, Proceedings of the Second InternationalVascular Symposium, 1986, abstract No. $20.3. 23. Hill SL, Martin D, Evans P. Massive vein thrombosis of the extremities. Am J Surg 1989;158:131-6. 24. Alkjaersig N, Fletcher AP, Sherry S. The mechanism of clot dissolution by plasmin. J Clin Invest 1959;38:1086-91. 25. Molina JE, Hunter D, Yedl~ckaJW. Thrombolytic therapy for iliofemoral venous thrombosis. Vascular Surgery 1992;26: 630-7. 26. Rutherford RB. The role of thrombectomy in the management of iliofemoral venous thrombosis. In: Rutherford RB, ed. Vascular surgery. 3rd ed. WB Saunders, Philadelphia; 1989:1569-74. 27. Dale WA. Crossover grafts iliofemoral venous occlusion. In: Bergan JJ, Yao JST, eds. Venous problems. Chicago: Yearbook Medical Publishers;1978:411-20. 28. Gruss J. Venous bypass for chronic venous insufficiency. In: Bergan JJ, Yao JST, eds. Venous disorders. Philadelphia: WB Saunders Co;1991:316-30.

Submitted Jan. 10, 1994; accepted March 7, 1994.