Endovascular Recanalization of the Thrombosed Filter-bearing Inferior Vena Cava Suresh Vedantham, MD, Thomas M. Vesely, MD, Naveen Parti, MD, Michael D. Darcy, MD, Thomas K. Pilgram, PhD, Gregorio A. Sicard, MD, and Daniel Picus, MD
PURPOSE: To evaluate the authors’ preliminary experience with use of endovascular methods to treat inferior vena cava (IVC) thrombosis in patients with IVC filters. MATERIALS AND METHODS: Catheter-directed thrombolysis, balloon maceration, mechanical thrombectomy, and stent placement were used to treat 10 patients with thrombosis of filter-bearing IVCs causing symptoms in 18 limbs. Procedural challenges, technical and clinical success, complications, postprocedural filter status, and postprocedural pulmonary embolism (PE) prophylaxis were monitored. RESULTS: Technical and clinical success were achieved in 15 of 18 (83%) and 14 of 18 symptomatic limbs (78%), respectively. Major bleeding (muscular hematoma) occurred in one patient (10%). Postprocedural PE prophylaxis included anticoagulation (n ⴝ 8) and placement of a new filter into a newly placed Wallstent (n ⴝ 1). During clinical follow-up, no clinically detectable PE was observed. Data pertaining to late limb status were available at a median of 19 months (range 1– 46 months) follow-up in seven patients: three patients were asymptomatic, two patients had ambulatory edema only, one patient had constant mild edema, and one patient had constant severe edema. Postprocedural filter stability was radiographically documented at a median of 255 days (range, 4 –1021 d) of follow-up. CONCLUSION: Endovascular recanalization of the occluded IVC is feasible even in the presence of an IVC filter. J Vasc Interv Radiol 2003; 14:893–903 Abbreviations: DVT ⫽ deep venous thrombosis, IVC ⫽ inferior vena cava, MT ⫽ mechanical thrombectomy, PE ⫽ pulmonary embolism, tPA ⫽ tissue plasminogen activator
INFERIOR vena cava (IVC) thrombosis is a well-recognized complication of IVC filter placement, occurring in 2%–10% of treated patients (1). Afflicted patients initially experience lower extremity and/or body wall edema and pain. Depending on collateral formation and the extent of venous involvement and valvular dam-
From the Mallinckrodt Institute of Radiology (S.V., T.M.V., N.P., M.D.D., T.K.P., D.P.) and Department of Surgery (G.A.S.), Washington University School of Medicine, 510 South Kingshighway Boulevard, Box 8131, St. Louis, Missouri 63110. Received November 11, 2002; revision requested January 12; revision received and accepted February 11. Address correspondence to S.V., E-mail: [email protected]
None of the authors have identified a potential conflict of interest. © SIR, 2003 DOI: 10.1097/01.RVI.0000083842.97061.c9
age, the long-term sequelae of IVC thrombosis can range from mild ambulatory lower extremity swelling to incapacitating edema at rest, venous claudication, and/or venous ulcers. Anticoagulation may partly ameliorate the acute symptomatology and also decreases the subsequent risk of pulmonary embolism (PE). However, thrombus progression occurs in many patients, and anticoagulation is unlikely to prevent severe postphlebitic symptoms from eventually developing (2– 4). When performed in carefully selected patients with iliofemoral deep venous thrombosis (DVT) by surgeons with specialized venous expertise, venous thrombectomy and venous bypass are associated with 75%– 85% and 60%– 65% 3-year patency rates, respectively (5–7). However, surgical therapy carries significant operative risks and morbidity, and re-
ports detailing its use in treating IVC thrombosis are extremely limited (8). Endovascular therapy has demonstrated short-term effectiveness in treating patients with iliofemoral DVT and limited numbers of patients with IVC thrombosis (9 –19). However, little attention has been paid to the unique challenges associated with endovascular recanalization of the thrombosed IVC containing a permanent filtration device. The purpose of this study was to evaluate safety and efficacy of endovascular recanalization of the thrombosed filter-bearing IVC.
MATERIALS AND METHODS Patients Our institutional review board approved this study. A computer search of our interventional radiology database was used to identify all patients
Endovascular Recanalization of the Filter-bearing IVC
Table 1 Demographics and Filter History in Patients Undergoing Endovascular IVC Recanalization Patient No.
PE despite anticoagulation PE despite anticoagulation
Bird’s Nest Greenfield
2 months 7 days
DVT and PE, preoperative pelvic mass resection DVT and PE, recent intracranial bleeding DVT and bleeding complication with anticoagulation DVT and PE, bleeding complication with anticoagulation DVT propagation despite anticoagulation DVT and PE, recent endovascular AAA repair DVT propagation and PE despite anticoagulation DVT and recent MVA/trauma
Bird’s Nest Bird’s Nest Bird’s Nest
Sympt Duration Acute: 3 days Acute: 7 days
Bilateral LE edema Bilateral LE edema and phlegmasia Bilateral LE edema
Bilateral LE edema
Acute: 6 days
Right LE edema and pain
Chronic: 56 months Chronic: 1 month
Acute: 2 days
Right LE edema
Bilateral LE edema and pain Bilateral LE edema
Chronic: 7 months Acute: 7 days
Bilateral LE edema and phlegmasia Bilateral LE edema and pain
Acute: 7 days
Chronic: 8 months
* Time interval between filter placement and the endovascular procedure. Note.—LE ⫽ lower extremity; MVA ⫽ motor vehicle accident.
who had IVC thrombosis identified during IVC venography from January 1996 through August 2002. These radiology reports were examined to identify patients who underwent attempted endovascular recanalization of the IVC. From this group, 10 patients were identified as having undergone attempted endovascular recanalization of filter-bearing IVCs in the treatment of 18 symptomatic limbs. This study was conducted based on these patients. Being a retrospective study, there were no prospectively established criteria for determining which patients with IVC thrombosis were offered endovascular therapy. Hence, the patients included in this study were not treated consecutively. However, each patient fulfilled three general criteria. First, each patient had venographyproven thrombosis of a filter-bearing IVC. In nine patients, IVC thrombosis involved the filter-bearing infrarenal IVC segment and at least one iliofemoral venous system. In one patient, isolated suprarenal IVC thrombosis was present distant from an infrarenal filter. Second, each patient underwent initial evaluation by a vascular surgeon and was subsequently referred for endovascular therapy. Third, each
patient had experienced a recent increase in symptom severity despite conservative therapy and had lifestyle-limiting lower extremity symptoms at the time of treatment. Specifically, all 10 patients reported severe lower extremity swelling. Two patients (four limbs) were treated for phlegmasia and two other patients had massive lower extremity swelling that prevented ambulation. Three patients also experienced chronic lower extremity pain. The patients included five women and five men with a median age of 58 years (range, 31–75 y). Six patients with acute IVC thrombosis (symptom duration ⱕ10 days) had symptoms for a median of 6.5 days (range, 2–7 d). Four patients with chronic IVC thrombosis (symptom duration ⱖ10 days) had symptoms for a median of 7.5 months (range, 1–56 months) (9). Four patients had known malignancies and seven patients had a history of PE. Filters Each of the 10 patients in this study had an infrarenal IVC filter in place at the time of endovascular treatment. The filters had been placed in our institution in five patients and in other
institutions in five patients. Patient demographics, symptoms, indication for filter placement, filter type, and the interval between filter placement and the endovascular procedure for each patient are recorded in Table 1. Endovascular Procedures Informed consent was obtained from each patient after a discussion of the risks and benefits of off-label use of catheter-directed thrombolysis, mechanical thrombectomy (MT), venous stent placement, and surgical treatment alternatives. Declaration of Helsinki principles were adhered to in all cases. In this retrospective review, the specific method for performing endovascular recanalization of the IVC was at the judgment of the treating interventionalist in each individual patient. Approaches were performed via ultrasound (US)-guided entry into the right internal jugular vein (n ⫽ 4), right popliteal vein (n ⫽ 4), left popliteal vein (n ⫽ 4), right common femoral vein (n ⫽ 1), left common femoral vein (n ⫽ 1), right posterior tibial vein (n ⫽ 1), and left anterior tibial vein (n ⫽ 1). There was a median of 1.6 access sites per patient (range, 1–3). The high fre-
quency of popliteal and infrapopliteal venous access reflected the fact that nine of the 10 patients had thrombus extension into at least one iliofemoral venous system. A multipurpose angiographic catheter was used to perform preliminary venography to define the location and extent of venous thrombosis, and a vascular sheath was subsequently inserted. Endovascular Thrombus Removal Methods Nine patients were initially treated with catheter-directed thrombolysis. In these patients, a multiple–side hole infusion catheter/wire system was positioned within the thrombosed venous segment. The thrombolytic agent was infused directly into the thrombus, as is standard for catheter-directed venous thrombolysis (9). The dosing and infusion parameters for administration of the thrombolytic agent given were dependent on physician preference and the extent of thrombus. One of three thrombolytic agents was administered: (i) urokinase (Abbott Laboratories, North Chicago, IL) was used in four patients (seven symptomatic limbs), with initial doses of 120,000 –250,000 U/h (median, 162,500 U/h; 100,000 U/h per limb); (ii) tissue plasminogen activator (tPA, Genentech, South San Francisco, CA) was used in one patient (one symptomatic limb), with an initial dose of 2.5 mg/h (weight-based regimen of 0.05 mg/kg/h); and (iii) reteplase (Centocor, Malvern, PA) was used in four patients (eight symptomatic limbs), with initial doses of 0.5–1.5 U/h (median, 0.88 U/h; 0.44 U/h per limb). All patients receiving urokinase and tPA underwent full heparinization to achieve partial thromboplastin times of 60 – 80 seconds. All patients receiving reteplase received subtherapeutic doses of heparin (300 – 400 U/h). Partial thromboplastin times and fibrinogen levels were monitored at 6-hour intervals, with dose reductions made when fibrinogen levels decreased below 150 mg/dL. Patients were given around-the-clock prophylactic intravenous cefazolin while the catheter was in place. Repeat venography was performed every 6 –12 hours to monitor the progress of thrombolysis. If residual thrombus was present, balloon macer-
Vedantham et al
ation was performed to encourage greater thrombolytic efficiency. MT devices were used in seven patients. In one patient (two symptomatic limbs), MT was performed without accompanying pharmacologic thrombolysis in the primary treatment of an acutely thrombosed IVC in a patient with a recent hemorrhagic stroke. In the remaining six patients, MT was used in conjunction with pharmacologic thrombolysis. The MT devices were used in an off-label fashion in the manner described in the manufacturers’ instructions. Devices used included the Amplatz Thrombectomy Device (Microvena, White Bear Lake, MN; n ⫽ 6), AngioJet (Possis, Minneapolis, MN; n ⫽ 2), and Oasis (Boston Scientific/Medi-tech, Natick, MA; n ⫽ 1). In two patients, more than one device was used. Typically, the device was advanced into the venous system through a 7–9-F vascular sheath. A guiding catheter was used for directional control of an MT device in one patient. In most cases, two back-andforth passes through the thrombosed venous segment (including the filterbearing segment) were made with use of the MT device. Balloon Venoplasty and Endovascular Stent Placement In most patients, thrombolytic infusion was continued until adequate thrombolysis was achieved, until no evidence of continued thrombolysis was seen on a follow-up venogram compared with the preceding venogram, or until 2 days of treatment had passed. At this point, residual iliac vein and/or IVC thrombus or stenosis was treated with balloon venoplasty and endovascular stent placement (with the exception of one patient who was treated before we began to use stents in the venous system). In total, seven patients received stents. The stents were placed in the IVC and both iliac veins (n ⫽ 4), in the IVC and one iliac vein (n ⫽ 2), or in the external iliac vein only (n ⫽ 1). In these seven patients, a median of five stents (range, 1–16) were placed. In three patients, stents were placed through the IVC filters to above the highest extent of thrombus. In two additional patients, the upper end of the stents was within the filter-bearing IVC segment. The stents used were 10 –20-mm self-
expandable Wallstents (Boston Scientific/Medi-tech; n ⫽ 7), Palmaz 308 balloon-expandable stents (Johnson & Johnson, Warren, NJ; n ⫽ 1), and 14-mm self-expandable Smart stents (Cordis Endovascular, Miami, FL; n ⫽ 1). Residual femoral vein thrombus or stenosis was treated with balloon venoplasty. Patients underwent systemic anticoagulation with oral warfarin to a target International Normalized Ratio of 2–3 after the procedure when possible, and those receiving stents were also administered indefinite oral antiplatelet therapy (aspirin) if there were no contraindications. Venographic Evaluation The venograms of each patient were retrospectively reviewed. In each patient, the overall degree of thrombolysis was categorized as complete, partial, or none according to the method employed in a large venous thrombolysis registry study (12). The venograms from each patient were evaluated to determine if technical success— defined as restoration of patency in the entire above-knee thrombosed venous segment in a limb with residual stenosis/occlusion of less than 30% diameter on the final venogram—was achieved. Clinical Evaluation The patients’ medical records, radiology reports, and procedural data were reviewed. The endovascular treatment modalities used, vascular access sites used, technical challenges encountered, thrombolytic infusion time, total thrombolytic agent dose, end-of-procedure filter appearance, clinical success, complications, and method of postprocedural PE prophylaxis were recorded. Clinical success was defined as considerable improvement in lower extremity swelling and/or pain that lasted at least 3 days or until hospital discharge, in conjunction with limb salvage (20). Major bleeding was defined as intracranial bleeding or bleeding resulting in death, transfusion, surgery, or cessation of thrombolytic therapy, per Society of Interventional Radiology (SIR) reporting standards (21). Other complications were classified as major and minor per SIR reporting standards (22), and postprocedural renal failure
Endovascular Recanalization of the Filter-bearing IVC
Table 2 Results of Endovascular IVC Recanalization Procedures Patient No. Lytic Agent
No. Follow-up of Technical Clinical Interval Degree of Lysis Stents Success Success Complications (mo)
1 2 3 4 5 6
Urokinase Urokinase Urokinase None Urokinase Alteplase
None None ATD AngioJet, ATD AngioJet Oasis, ATD
Partial Complete Complete Partial Partial Partial
0 1 0 0 5 5
0/2 2/2 2/2 2/2 1/1 0/1
0/2 2/2 2/2 2/2 1/1 0/1
None Partial Partial
16 7 5
2/2 2/2 2/2
2/2 2/2 1/2
None None None None None Muscular bleeding None None None
7 8 9
Reteplase Reteplase Reteplase
ATD ATD ATD
Late Limb Status
26 2 8 1 46 42
Constant severe edema Unknown Unknown Constant mild edema Asymptomatic Asymptomatic
1 19 5
Ambulatory edema Ambulatory edema Left BKA, Right unknown Asymptomatic
Note.—ATD ⫽ Amplatz Thrombectomy Device; BKA ⫽ below– knee amputation.
was defined as a 20% or greater increase in serum creatinine level after the procedure per SIR reporting standards (23). Clinical follow-up in each patient was directed at evaluating three areas of interest: (i) information pertaining to the long-term status of the treated limbs was acquired via telephone follow-up with the patients, supplemented by medical record review and telephone follow-up with the referring physicians; (ii) although we did not systematically obtain imaging follow-up to evaluate filter position, information pertaining to the long-term status of the filters was obtained by review of medical records and postprocedural abdominal imaging studies performed for other indications; and (iii) the incidence of postprocedural clinically detectable PE (defined as having occurred in any patient with a pulmonary arteriogram showing PE, a high-probability ventilation/perfusion scan, or recurrent unexplained episodes of dyspnea during followup) was assessed during telephone follow-up with the patients and their physicians, supplemented by review of medical records and radiology reports. Objective imaging follow-up to assess for recurrent DVT and the occurrence of PE was not routinely performed unless clinically indicated. In patients who were lost to more recent follow-up, the clinical follow-up interval was defined as the time between the IVC recanalization procedure and the last time point at which informa-
tion about the clinical issue in question (eg, lower extremity symptom status or incidence of clinical PE) was clearly documented in the medical record. Data Analysis and Statistics The frequency of complete thrombolysis, partial thrombolysis, major bleeding, minor and major complications, technical success, clinical success, postprocedural filter migration, and postprocedural clinically detectable PE were expressed as percentage values. The median thrombolytic infusion time and median total thrombolytic agent dose were also calculated. Calculations were performed with use of a JMP statistical software package (SAS Institute, Cary, NC).
RESULTS Initial Success and Complications Details pertaining to the endovascular procedures and the initial results of treatment in each patient are presented in Table 2. Technical success was achieved in 15 of 18 symptomatic limbs (83%), with no residual IVC stenosis in eight of 10 patients (80%). Clinical success was achieved in 14 of 18 symptomatic limbs (78%) as described later. Complete thrombolysis was observed in two patients (20%) with acute IVC thrombosis, with subsequent symptomatic relief (Fig 1). Partial thrombolysis was observed in
seven patients (70%), of whom four were successfully treated with endovascular stent placement. A fifth patient presented with severe bilateral phlegmasia. In this patient, venous patency and technical success were successfully achieved bilaterally and in the IVC, and the left lower extremity was successfully salvaged with use of thrombolysis and stents. However, the patient ultimately required belowknee amputation of the right leg, and did not therefore fulfill criteria for clinical success in that limb. The other two patients experiencing partial thrombolysis did not receive stents. One of these two patients had minimal residual IVC thrombus after thrombolysis and experienced symptomatic relief. The other patient had significant residual IVC thrombus but did not undergo stent placement because he was treated before we began to use stents in the venous system. This patient did not experience technical or clinical success in either limb. No thrombolysis was observed in one patient (10%) with chronic IVC thrombosis. This patient was subsequently successfully treated with iliocaval stents. One patient (10%) experienced a major bleeding complication that ultimately resulted in early IVC reocclusion and clinical failure (Fig 2). This patient experienced partial thrombolysis after tPA infusion at a rate of 2.5 mg/h, but we chose not to extend stents above the residual filter-adherent thrombus. Hence, technical suc-
Vedantham et al
cess was not achieved. This patient developed a pectineus muscle hematoma 2 days after thrombolytic treatment was discontinued, necessitating blood transfusion. No other major or minor complications were observed. Specifically, no patient experienced intracranial or gastrointestinal hemorrhage, clinically detectable PE, or renal failure. The median thrombolytic infusion time per patient was 25 hours (range, 6 –59 h). The median total doses per patient were 3.93 million U of urokinase (range, 2.35–7.38 million U), 15.0 mg tPA, and 23.9 U reteplase (range, 12.0 –31.5 U). The median total doses per treated limb were 2.75 million U urokinase (range, 1.18 –3.69 million U), 15.0 mg tPA, and 12.0 U reteplase (range, 6.0 –15.8 U). Lower Extremity Clinical Follow-up Follow-up information about the postprocedural lower extremity symptom status was available in seven of the 10 patients at a median clinical follow-up interval of 19 months (range, 1– 46 months) after the IVC recanalization procedures. Three patients currently have no lower extremity pain or swelling, including one patient in whom technical or clinical success was not initially achieved. Two patients reported the absence of lower extremity edema when resting, but reported late-day edema on days with greater than usual ambulatory activity; in both these patients, postprocedural symptoms were greatly improved over their preprocedural status. Two patients reported chronic lower extremity edema; the one with more severe symptoms had not initially experienced technical or clinical success. No lower extremity ulcerations were reported, and all three patients initially presenting with chronic lower extremity pain reported successful resolution of this symptom. Postprocedure Filter Status
Figure 1. Successful catheter-directed thrombolysis of the occluded IVC in a 64-year-old woman with bronchogenic carcinoma and a Greenfield filter. (a) Digital subtraction venography demonstrates extensive thrombosis of the Greenfield filter– bearing infrarenal IVC with patency of the suprarenal IVC. (b) The left iliac vein is also thrombosed. (c) After catheter-directed thrombolysis with urokinase and placement of a left external iliac vein stent (arrows), the left iliac vein and infrarenal IVC are patent with minimal residual thrombus. (d) The right iliac vein is also patent and the filter is unchanged in position.
Seven patients underwent aggressive balloon angioplasty, MT, and/or stent placement within their IVC filters, presumably rendering the filters nonfunctional. This includes three patients in whom stents were deployed through the filters to above the highest
Endovascular Recanalization of the Filter-bearing IVC
Figure 2. Failure to achieve technical or initial clinical success with endovascular intervention in a 71-year-old man with right lower extremity swelling for a duration of 22 days after radical prostatectomy and perioperative VenaTech filter placement. (a) Digital subtraction venogram confirms thrombosis of the filter-bearing IVC. (b) The thrombus extends through the right common and external iliac veins. (c) Catheter-directed thrombolysis with alteplase, MT with the Amplatz Thrombectomy Device, and balloon maceration were performed in sequence. The iliocaval venous segments are improved but significant residual wall-adherent thrombus remains, causing diffuse stenosis. In addition, bulky thrombus at the filter apex (arrows) was resistant to further maceration with use of angioplasty balloons and MT. (d) Wallstents 10 –14 mm in diameter were placed from the right external iliac vein into (but not through) the filter-bearing IVC segment. Completion venography demonstrates patency of the right common femoral vein and stent-implanted right external iliac vein. (e) The stent-implanted right common iliac vein and lower IVC are also patent, but persistent partial occlusion of the IVC is present at the level of the filter apex (arrows). Rethrombosis occurred 2 days later after a bleeding event necessitated discontinuation of anticoagulation.
extent of thrombus. Radiographs taken near the end of the endovascular procedures demonstrated deformation of the filters in four patients. Outright fracture of filter components was observed radiographically in one patient. No instances of clinically apparent caval wall penetration were reported. Postprocedural abdominal computed tomography (CT) scans were available in only one patient and did not demonstrate evidence of caval penetration. Abdominal radiographs obtained in eight patients at a median interval of 255 days (range 4 –1,021 d) after the procedure showed no evidence of IVC filter migration.
Pulmonary Embolism Prophylaxis The single patient treated with MT alone had a clear contraindication to anticoagulation (hemorrhagic stroke) and did not receive postprocedural anticoagulation. The remaining nine patients underwent anticoagulation after the procedures. One patient who had experienced multiple previous episodes of PE despite therapeutic anticoagulation underwent successful recanalization with thrombolysis and placement of Wallstents in the IVC and both iliac veins. In this patient, a new VenaTech filter was placed within a newly placed caval Wallstent for continued PE prophylaxis (Fig 3).
Follow-up information about the incidence of clinically detectable PE was available in nine of the 10 patients at a median clinical follow-up interval of 8 months (range, 1– 46 months) after the IVC recanalization procedures. No patients experienced clinically detectable PE during the follow-up period.
DISCUSSION Catheter-directed thrombolysis has proved effective in reestablishing venous patency in patients with iliofemoral DVT, with early clinical success rates of 80%– 85% in patients with acute symptomatology (12). Although the long-term success of this approach
Vedantham et al
Figure 3. The “destroy-and-replace” method of IVC recanalization and continued caval filtration in a 39-year-old woman with acute IVC and bilateral iliofemoral venous thrombosis. The woman had a history of PE despite therapeutic anticoagulation with subsequent TrapEase filter placement. (a) Digital subtraction venography demonstrates thrombosis of the TrapEase filter– bearing infrarenal IVC. The thrombus involved both iliofemoral venous systems. (b) Overnight catheter-directed thrombolytic infusion of reteplase was performed with infusion catheters placed from bilateral popliteal veins, and adjunctive MT with use of the Amplatz Thrombectomy Device was performed the next day. Aggressive balloon venoplasty was performed below and within the IVC filter. Repeat venography (c) revealed significant persistent thrombus within the IVC filter and left iliac vein. (d) The right iliac vein also still contains significant thrombus. (e) A 20-mm Wallstent was deployed through the IVC filter from a right internal jugular vein approach. (f) Overlapping “kissing” 14-mm Wallstents were extended into both iliac veins. Venography after stent placement demonstrates a patent stentimplanted lower IVC and common iliac veins with minimal residual thrombus and patency of the IVC (g). Because of the patient’s history of PE despite therapeutic anticoagulation, a VenaTech filter was placed within the caval Wallstent (h). Note the fractured TrapEase filter (arrows).
Endovascular Recanalization of the Filter-bearing IVC
has yet to be demonstrated in prospective controlled trials, one retrospective study has documented improved health-related quality of life in patients with DVT undergoing catheter-directed thrombolytic therapy compared with patients receiving anticoagulation alone (14). However, bleeding complications during DVT thrombolysis have been reported to occur in 0%–25% patients in single-center series and occurred in 11% of patients in a multicenter registry study (9 –19). In addition, the relatively long infusion times (mean of 48 hours in the registry study) often required to treat DVT can be difficult to tolerate for some patients, and complications may become more frequent with longer infusion durations (12,16). Because of the greater thrombus burden and the frequent involvement of bilateral iliofemoral veins, patients with IVC thrombosis often require even longer infusions and greater thrombolytic doses than the average patient with iliofemoral DVT, potentially increasing the risks, discomfort, and costs of therapy. For these reasons, it is important to determine whether patients with particularly challenging anatomic considerations, such as the presence of a filter, should be considered candidates for endovascular therapy. It would also be important to know whether anatomic and functional disruption of the filter during endovascular manipulations is associated with longterm clinical consequences. The existing literature indicates that relatively few patients with IVC thrombosis have been treated with use of endovascular means. Several published iliofemoral DVT thrombolysis series have included a few patients who had IVC involvement (15–19). Only two studies have included larger numbers of patients with IVC involvement. In the initial Stanford University experience with catheter-directed venous thrombolysis, 16 of 41 treated patients had IVC involvement (11). Similarly, in the University of Minnesota experience, 17 of 77 treated patients had IVC thrombosis (10). However, in neither of these series were the results of the IVC thrombosis subset reported separately from those of patients with thrombosis limited to the iliofemoral venous system. To our knowledge, the largest series in which the results of thrombolytic therapy for IVC thrombosis are separately reported is that of
Angle et al (24), in which technical success of thrombolytic therapy and adjunctive balloon angioplasty was described in seven of eight patients with IVC thrombosis of 2–21-day duration. The published experience reporting endovascular therapy for patients with filter-bearing venae cavae is even more sparse. Because the majority of patients with IVC filters have relative or absolute contraindications to anticoagulation therapy, only a minority of patients with filter thrombosis are considered suitable candidates for endovascular therapy. In the IVC thrombolysis series of Angle et al (24), the only patient not experiencing technical success was one of three patients with IVC filters. However, all three patients with filters did experience clinical improvement. Hansen et al (25) described the use of catheter-directed thrombolysis to successfully treat one patient with filter-related thrombosis, and Tarry et al (26) reported the successful treatment of three patients with filter-related IVC thrombosis. In the report of Hansen et al (25), pulsespray thrombolysis was used to address residual caval thrombus that was present on the filter after overnight thrombolytic infusion; although clinical success was achieved, the final venogram did demonstrate significant residual thrombus on the filter. Finally, Ryu et al (27) used catheter-directed thrombolytic therapy in a patient with filter-related thrombosis but observed a severe bleeding complication mandating cessation of therapy. These authors cited the lack of proven long-term benefit of catheter-directed thrombolysis and the potential for complications as reasons to exercise significant caution in selecting patients for endovascular therapy. The current study, although small in terms of number of patients, is the largest reported experience in which endovascular therapy was used to treat IVC thrombosis in filter-bearing cavae. In our experience, endovascular methods were safe and effective in the short term in this clinical setting, with a procedural success rate comparable to that reported for iliofemoral DVT therapy. We did observe major bleeding in one of 10 patients (10%); although this is comparable to major bleeding rates reported in other DVT thrombolysis studies, it is possible that
our use of a weight-based tPA regimen in this patient contributed to this bleeding event (12). When using tPA, we currently use a non–weight-based regimen of 0.5–1.0 mg/h. Like previous investigators, we observed that the presence of an IVC filter can present significant technical challenges to achieving endovascular patency. Several major issues are worthy of consideration when planning endovascular therapy in patients with IVC filter thrombosis. First, the presence of the filter presents specific challenges in terms of guide wire recanalization and the passage of endovascular catheters and devices. In the subset of patients with chronic symptomatology, simple catheter/wire manipulations may not be sufficient to cross the tightest part of the occluded venous segment, which is usually the filter-bearing segment. In three patients in our series, difficulty in traversing the filter-bearing IVC segment was the primary reason why an additional access site had to be used. In one of these three patients, antegrade guide wire access across the filter-bearing IVC segment could not be obtained, but the segment was successfully bridged from an internal jugular vein approach. In the other two patients, a guide wire was successfully manipulated across the occlusion, but a catheter would not follow past the filter-bearing IVC segment. In these cases, an Amplatz goose neck snare (Microvena) was used to capture the guide wire from an additional internal jugular vein access; by applying traction to the through-and-through guide wire, catheters could finally be passed through the filter-bearing segment. The long distance from the popliteal vein access sites to the IVC presented two distinct types of difficulty. First, when an occlusion cannot be crossed with preliminary catheter/ wire manipulations, a long sheath is usually used to provide additional support for subsequent guide wire recanalization attempts. However, during one such procedure, we did not have extra-long guiding catheters/ sheaths to provide optimal support for our attempts to cross the occluded segment. Second, in two patients, the working length of our preferred stent was not sufficient to reach the uppermost aspect of the IVC occlusion via the popliteal vein; in one of these pa-
tients, a different stent was used, and in the other patient, jugular venous access was used to place the uppermost caval stent. Similarly, in two cases, jugular access was required to permit the use of MT devices that were too short to reach the IVC from the popliteal vein. In patients with chronic occlusion, even when infusion catheter placement was successful, aggressive balloon venoplasty was often required to enlarge the lumen sufficiently to permit passage of MT devices and endovascular stents. The use of high-pressure balloons resistant to rupture is recommended when dilating the filterbearing segment. In two of our patients, balloon venoplasty within the filter-bearing segments resulted in balloon rupture, and in one case, the deflated balloon became entangled with the filter and was removed with considerable difficulty. In four cases, deformation of the filter during balloon venoplasty and/or stent placement was observed fluoroscopically, and in one case, frank fracture of filter components was visualized. No clinical sequelae of filter damage was observed in these patients, but the patient with fractured filter components was later lost to follow-up. When performing venous interventions for DVT, it is optimal to access the venous system below the lowest extent of thrombus to adequately ensure suitable inflow to the treated segment. However, in cases in which the thrombosed segment encompasses both the infrarenal IVC and the entire iliofemoral venous system to the knee level, access to the proximal and distal aspects of the occlusion may not be possible from a single venous entry site, necessitating the simultaneous use of internal jugular vein and popliteal/tibial vein access sites. The simultaneous use of both access sites may be required to obtain through-andthrough access; however, we must state that working in this fashion can be tiresome to the interventionalist and the patient being asked to assume nonstandard positions on the procedure table. For these reasons, we believe that, when bilateral lower-extremity symptoms are present, preprocedural imaging characterization of the upper and lower extent of the thrombus (either by contrast-enhanced CT or by a combination of US
Vedantham et al
and venography) is very helpful in determining what access site(s) are needed to best perform the procedure. The second major issue regards the creation of a sufficient flow lumen within the filter-bearing segment. The cohort of patients with IVC filter thrombosis may have an inherently higher likelihood of being hypercoagulable, which may adversely impact thrombus removal. We observed more residual thrombus within the filterbearing segment of the IVC after thrombolysis. Although there could be several possible explanations for this finding, we believe that the process by which IVC filter thrombosis occurs involves embolization of chronic thrombus into the filter from lower extremity or pelvic sources, followed by acute thrombus deposition when flow is slowed sufficiently. It seems likely that the chronic component of the caval thrombus may be relatively resistant to complete thrombolysis, as is often observed with older thrombus. In addition, the presence of filter struts may render parts of the vessel inaccessible to MT devices and angioplasty balloons, preventing adequate thrombus maceration. Several approaches can be used to attempt to improve thrombus removal from the IVC and its filter-bearing segment. First, pulse-spray thrombolysis has proved to be an effective method in improving the efficiency of acute clot lysis (28). In one of our patients and in a previously described patient (25), pulse-spray infusion of 250,000 U of urokinase over a period of 30 minutes proved effective in removing residual acute thrombus in the filterbearing segment. Disadvantages of this approach include the need to give higher thrombolytic doses and the extra on-table procedure time involved. Second, we have found adjunctive MT with use of the Amplatz Thrombectomy Device to be very helpful in accelerating the removal of acute thrombus without adding to bleeding risks. Several studies (20,29 –33) have reported the successful use of MT devices in treating acute iliocaval DVT, both with and without concomitant pharmacologic thrombolysis. Potential benefits of MT devices include dose reduction, more rapid restoration of a flow channel through the thrombosed venous segment, and shortened overall thrombolytic infusion times (20,34).
The disadvantages of this approach are increased on-table procedure time, inability to adequately treat chronic thrombus with currently available devices, inability to access thrombus trapped behind filter struts, and PE if using nonaspirating devices without concomitant thrombolytic agents (29,31). Third, combination pharmacologic therapy with parenteral glycoprotein IIb/IIIa antagonists has shown initial promise in improving thrombolytic efficiency for peripheral arterial applications and could theoretically improve thrombolysis of resistant venous thrombus (35). However, the efficacy of this approach and the additional bleeding risk incurred (if any) have not yet been characterized, and we have not yet adopted this approach. Despite the use of these adjuncts to thrombolysis, many patients treated for IVC filter thrombosis will likely have significant residual thrombus after thrombolysis, particularly those with chronic symptomatology (12). Different strategies have been employed to treat these patients, but the results do not favor any particular management protocol. We and others have observed patients in whom partial thrombolysis with subsequent anticoagulation was sufficient to produce clinical improvement (24,25). However, one patient in the current series did develop early occlusion when residual thrombus was left within the filter apex after a bleeding complication that prompted cessation of anticoagulation. In patients with iliofemoral DVT, the 1-year primary patency rate of iliac vein stents in patients without malignancies has been reported to be more than 90% (11). For these reasons, when significant residual filter-adherent thrombus is present, we favor deployment of stents through the filter to above the highest level of thrombus. By providing the most continuous and uninterrupted luminal surface, stents might help to minimize the risk of rethrombosis in patients who later require discontinuation of anticoagulation (as we observed in one patient) or who have periods when the degree of anticoagulation is subtherapeutic. Recently, some investigators have been using primary stent placement to recanalize chronically occluded venous segments, reserving preceding
Endovascular Recanalization of the Filter-bearing IVC
thrombolysis for those patients with evidence of acute thrombus formation. Entwisle et al (36) reported on three patients in whom the Wallstent endoprosthesis was used to treat malignant IVC occlusion without thrombolytic therapy. In a larger experience, Razavi et al (37) treated 17 patients with primary stent placement for chronic IVC occlusions; five patients who had acute thrombus formation received preceding thrombolytic therapy. Technical success was achieved in 15 of 17 patients, with a primary patency of 80% at a mean follow-up of 19 months and no procedure-related complications. We concur that aggressive stent placement will usually be the best endovascular solution for patients with chronic IVC thrombosis. However, the use of preceding thrombolytic infusion to “soften” the thrombus may enable a larger venous diameter to be achieved with stent placement. Therefore, our current methods include an initial relatively short course of pharmacologic thrombolysis in patients with chronic iliocaval thrombosis and no contraindications. The third major issue is the status of the filter itself after the completion of endovascular therapy. Any interventions within the IVC filter can predispose to filter migration or caval wall penetration, particularly recently placed filters that have not yet incorporated into the caval wall (1,27). Fortunately, despite a very aggressive approach to intervention within the filter-bearing segment—we significantly damaged several filters with balloon venoplasty and we fractured one filter—we did not observe these complications. Although caution is urged when performing endovascular manipulations in the presence of a newly placed IVC filter, the experience reported in this study suggests that the presence of a filter should not in itself deter the interventionalist from attempting endovascular therapy in the symptomatic patient with IVC thrombosis. Finally, several factors can influence subsequent decisions regarding the optimal method of future PE prophylaxis in patients with filters rendered nonfunctional by intervention. In most patients, anticoagulation may be sufficient to prevent PE. However, in patients with proven PE despite therapeutic anticoagulation, signifi-
cant residual DVT above the popliteal vein level, or contraindications to long-term anticoagulation, continued caval filtration is indicated. If the initial filter placement was sufficiently low, it may be possible to deploy a second filter within the infrarenal IVC above the first filter. If there is not sufficient space below the renal veins to accommodate a second filter, the situation becomes more difficult. We are reluctant to place suprarenal filters in patients whose initial IVC filters have developed thrombosis; therefore, in the one patient requiring continued filtration, we placed the new filter within a caval Wallstent. We postulate that this “destroy-and-replace” method of recanalizing the occluded filter-bearing IVC and providing continuing infrarenal caval filtration may be safer than resorting to suprarenal filter placement in this patient subset. However, a larger experience and longer follow-up would be required to determine if this is indeed the case. The most significant limitations of this retrospective study are the small patient number and the fact that our endovascular technique evolved during the 6-year time period. An objective symptom assessment instrument was not used to assess clinical success, and the duration of clinical follow-up was fairly short in some cases. Objective imaging follow-up was not used to definitively evaluate for IVC rethrombosis, filter penetration, and subclinical postprocedural PE.
CONCLUSIONS Our experience and literature review have led us to several conclusions: (i) Endovascular recanalization of the occluded, filter-bearing IVC is technically feasible and was reasonably safe in our preliminary experience. Many patients with contraindications to anticoagulation at the time of filter placement later become reasonable candidates for thrombolysisbased endovascular therapy. (ii) The presence of an IVC filter should not necessarily dissuade the interventionalist from attempting endovascular treatment. The presence of the IVC filter presents technical challenges that can be overcome with use of multiple vascular access sites and advanced interventional techniques. (iii) In most cases, catheter-directed thrombolysis
will not produce complete thrombolysis within the filter-bearing segment. While this does not in itself preclude a clinically successful result, we currently favor the placement of stents through the IVC filter when residual filter-adherent thrombus is present. (iv) IVC filters damaged during endovascular repair must be presumed nonfunctional but have been stable in our preliminary experience. (v) In our experience, continued anticoagulation provided effective postprocedural PE prophylaxis. In those situations in which continued filtration is indicated, a new filter can be placed into the stent-implanted IVC segment if sufficient space is not present in the infrarenal IVC. References 1. Grassi CJ, Swan TL, Cardella JF, et al. Quality improvement guidelines for percutaneous permanent inferior vena cava filter placement for the prevention of pulmonary embolism. J Vasc Interv Radiol 2001; 12:137–141. 2. Hirsh J, Hoak J. Management of deep vein thrombosis and pulmonary embolism: a statement for healthcare professionals. Circulation 1996; 93:2212–2245. 3. Strandness DE, Langlois Y, Cramer M, Randlett A, Thiele BL. Long-term sequelae of acute venous thrombosis. JAMA 1983; 250:1289 –1292. 4. Prandoni P, Lensing AWA, Cogo A, et al. The long-term clinical course of acute deep-vein thrombosis. Ann Intern Med 1996; 125:1–7. 5. Eklof B, Kistner RL. Is there a role for thrombectomy in iliofemoral venous thrombosis? Semin Vasc Surg 1996; 9:34 – 45. 6. Jost CJ, Gloviczki P, Cherry KJ, et al. Surgical reconstruction of iliofemoral veins and the inferior vena cava for nonmalignant occlusive disease. J Vasc Surg 2001; 33:320 –328. 7. Juhan CM, Alimi YS, Barthelemy PJ, Fabre DF, Riviere CS. Late results of iliofemoral venous thrombectomy. J Vasc Surg 1997; 25:417– 422. 8. Neglen P, Al-Hassan HK, Endrys J, et al. Iliofemoral venous thrombectomy followed by percutaneous closure of the temporary arteriovenous fistula. Surgery 1991; 110:493– 499. 9. Semba CP, Dake MD. Iliofemoral deep venous thrombosis: aggressive therapy with catheter-directed thrombolysis. Radiology 1994; 191:487– 494. 10. Bjarnasson H, Kruse JR, Asinger DA, et al. Iliofemoral deep venous thrombosis: safety and efficacy outcome during 5 years of catheter-directed thrombo-
lytic therapy. J Vasc Interv Radiol 1997; 8:405– 418. Semba CP, Dake MD. Catheter-directed thrombolysis for iliofemoral venous thrombosis. Semin Vasc Surg 1996; 9:26 –33. Mewissen MW, Seabrook GR, Meissner MH, Cynamon J, Labropoulos N, Haughton SH. Catheter-directed thrombolysis for lower extremity deep venous thrombosis: report of a national multicenter registry. Radiology 1999; 211:39 – 49. Grossman C, McPherson S. Safety and efficacy of catheter-directed thrombolysis for iliofemoral venous thrombosis. AJR Am J Roentgenol 1999; 172:667– 672. Comerota AJ, Throm RC, Mathias SD, Haughton S, Mewissen M. Catheterdirected thrombolysis for iliofemoral deep venous thrombosis improves health-related quality of life. J Vasc Surg 2000; 32:130 –137. Patel NH, Stookey KR, Ketcham DB, Cragg AH. Endovascular management of acute extensive iliofemoral deep venous thrombosis caused by May-Thurner syndrome. J Vasc Interv Radiol 2000; 11:1297–1302. Verhaeghe R, Stockx L, Lacroix H, Vermylen J, Baert AL. Catheter-directed lysis of iliofemoral vein thrombosis with use of rt-PA. Eur Radiol 1997; 7:996 –1001. Chang R, Cannon RO, Chen CC, et al. Daily catheter-directed single dosing of t-PA in treatment of acute deep venous thrombosis of the lower extremity. J Vasc Interv Radiol 2001; 12:247–252. Castaneda F, Li R, Young K, Swischuk JL, Smouse B, Brady T. Catheter-directed thrombolysis in deep vein thrombosis with use of reteplase: immediate results and complications from a pilot study. J Vasc Interv Radiol 2002; 13:577–580. Ouriel K, Katzen B, Mewissen M, et al. Reteplase in the treatment of peripheral arterial and venous occlusions: a
Vedantham et al
pilot study. J Vasc Interv Radiol 2000; 11:849 – 854. Vedantham S, Vesely TM, Parti N, Darcy M, Hovsepian DM, Picus D. Lower extremity venous thrombolysis with adjunctive mechanical thrombectomy. J Vasc Interv Radiol 2002; 13: 1001–1008. Patel N, Sacks D Patel RI, et al. SCVIR reporting standards for the treatment of acute limb ischemia with use of transluminal removal or arterial thrombus. J Vasc Interv Radiol 2001; 21:559 –570. Sacks D, Marinelli DL, Martin LG, et al. Reporting standards for clinical evaluation of new peripheral arterial revascularization devices. J Vasc Interv Radiol 1997; 8:137–149. Rundback JH, Sacks D, Kent KC, et al. Guidelines for the reporting of renal artery revascularization in clinical trials. J Vasc Interv Radiol 2002; 13:959 – 974. Angle JF, Matsumoto AH, Al Shammari M, et al. Transcatheter regional urokinase therapy in the management of inferior vena cava thrombosis. J Vasc Interv Radiol 1998; 9:917–925. Hansen ME, Miller GL III, Starks KC. Pulse-spray thrombolysis of inferior vena cava thrombosis complicating filter placement. Cardiovasc Intervent Radiol 1994; 17:38 – 40. Tarry WC, Makhoul RG, Tisnado J, Posner MP, Sobel M, Lee HM. Catheter-directed thrombolysis following vena cava filtration for severe deep venous thrombosis. Ann Vasc Surg 1994; 8:583–590. Ryu RK, Durham JD, Kumpe DA. Transcatheter venous thrombolysis— pitfalls and pratfalls: a case discussion of indications, technique, and alternatives. J Vasc Interv Radiol 1997; 8:1005– 1010. Bookstein JJ, Fellmeth B, Roberts A, Valji K, Davis G, Machado T. Pulsedspray pharmacomechanical thrombolysis: preliminary clinical results. AJR Am J Roentgenol 1989; 152:1097–1100.
29. Delomez M, Beregi J, Willoteaux S, et al. Mechanical thrombectomy in patients with deep venous thrombosis. Cardiovasc Intervent Radiol 2001; 24: 42– 48. 30. Gandini R, Maspes F, Sodani G, Masala S, Assegnati G, Simonetti G. Percutaneous ilio-caval thrombectomy with the Amplatz device: preliminary results. Eur Radiol 1999; 9:951–958. 31. Vorwerk D, Gunther RW, Wendt G, et al. Iliocaval stenosis and iliac venous thrombosis in retroperitoneal fibrosis: percutaneous treatment by use of hydrodynamic thrombectomy and stenting. Cardiovasc Intervent Radiol 1996; 19:40 – 42. 32. Uflacker R. Mechanical thrombectomy in acute and subacute thrombosis with use of the Amplatz device: arterial and venous applications. J Vasc Interv Radiol 1997; 8:923–932. 33. Kasirajan K, Gray B, Ouriel K. Percutaneous AngioJet thrombectomy in the management of extensive deep venous thrombosis. J Vasc Interv Radiol 2001; 12:179 –185. 34. Johnson SP, Cutts S, Durham JD, Krysl J, Kumpe DA. Initial experience with percutaneous mechanical thrombectomy using the Amplatz Thrombectomy Device for the treatment of deep venous thrombosis. J Vasc Interv Radiol 2000; 11(suppl):196 –197. 35. Tepe G, Schott U, Erley CM, et al. Platelet glycoprotein IIb/IIIa receptor antagonist used in conjunction with thrombolysis for peripheral arterial thrombosis. AJR Am J Roentgenol 1999; 172:1343–1346. 36. Entwisle KG, Watkinson FJ, Hibbert J, Adam A. The use of the Wallstent endovascular prosthesis in the treatment of malignant inferior vena cava obstruction. Clin Radiol 1995; 50:310 –313. 37. Razavi MK, Hansch EC, Kee ST, et al. Chronically occluded inferior vena cavae: endovascular treatment. Radiology 2000; 214:133–138.