|
|
||||||||
Vascular and Interventional Radiology |
1 Departments of Radiology (L.D.S., H.M.M., C.T.M., M.A.G., B.L.D.)
2 Vascular Medicine (M.G.G.), Cleveland Clinic Foundation, Hospital Hb-6, 9500 Euclid Ave, Cleveland, OH 44195.
| Abstract |
|---|
|
|
|---|
MATERIALS AND METHODS: Forty-one patients with acute upper extremity DVT and contraindications to or unsuccessful anticoagulation underwent percutaneous placement of a superior vena caval filter for prevention of PE. Four types of filters were used. Follow-up chest radiographs were used to detect filter migration, dislodgment, and fracture. Placements of central venous and Swan-Ganz catheters after filter insertion were recorded. Patients were followed up clinically for evidence of superior vena cava syndrome and PE. Kaplan-Meier survival rates were determined. Follow-up was 1 day to 221 weeks.
RESULTS: No complications such as filter migration, dislodgment, or fracture occurred (median follow-up, 12 weeks). No patients developed clinical evidence of PE due to upper extremity thrombosis or superior vena cava syndrome (median follow-up, 15 weeks). Catheters were placed subsequent to filter placement in 23 patients (56%) without complication.
CONCLUSION: Percutaneous filter placement in the superior vena cava is a safe and effective method for preventing symptomatic PE due to acute upper extremity DVT in patients in whom therapeutic anticoagulation has failed or is contraindicated.
Index terms: Embolism, pulmonary, 60.72 Veins, extremities Veins, thrombosis, 91.442, 91.751 Venae cavae, filters, 946.1267 Venae cavae, interventional procedure, 946.1267
| Introduction |
|---|
|
|
|---|
Partial mechanical interruption of the inferior vena cava (IVC) by means of percutaneous placement of a filter is a well-established method for preventing PE due to lower extremity DVT. Percutaneous placement of a filter in the superior vena cava (SVC) is a potentially attractive method of preventing PE due to upper extremity DVT. Such mechanical protection must, however, be shown to be safe and to protect against recurrent PE.
In 1985, Langham et al (7) reported that placement of the Greenfield filter in the SVC of 11 dogs was well tolerated at 3-month follow-up, protected against PE, and allowed thrombus resolution while caval patency was maintained. Despite this, there have been only 11 cases of SVC filter placement in humans reported in the English-language literature (813), to our knowledge. Although in each of these reports, the SVC filter was effective in preventing PE, the maximum patient follow-up was only 14 months (8).
Percutaneous filter insertion in the SVC is technically more demanding than insertion in the IVC because of the relatively small area for filter deployment. This could theoretically result in a higher complication rate for SVC filter insertion. Furthermore, recognized complications of IVC filter placement could potentially be more severe for SVC filter placement. Filter migration is more likely to result in an intracardiac position. Caval perforation may result in cardiac or aortic injury. SVC occlusion is more likely to result in substantial morbidity because of the reduction in potential collateral vessel pathways. In addition, the safety of central venous or Swan-Ganz catheter placement after SVC filter placement has yet to be determined.
We report our 9-year experience of SVC filter placement for PE prophylaxis in 41 patients with acute upper extremity DVT. The purpose of our study was to evaluate the safety and efficacy of percutaneous filter placement in the superior vena cava for prevention of PE from acute upper extremity DVT in patients with contraindications to, or failure of, anticoagulation.
| MATERIALS AND METHODS |
|---|
|
|
|---|
The patient population included 23 men and 18 women with a mean age of 61 years (age range, 3385 years). Chart review was used to determine the reason for hospital admission and the risk factors for upper extremity DVT, including the presence of a central venous catheter at the site of thrombosis. The presence of PE prior to filter placement, simultaneous or prior lower extremity DVT and IVC filter placement, and the indications for SVC filter placement were also documented. The extent of upper extremity DVT was determined from the upper extremity venogram, the duplex US study, or both, obtained before filter placement.
Superior vena cavograms were obtained in all patients prior to filter deployment to determine the caval diameter and to exclude venous anomalies, central venous stenosis, and SVC thrombus. The filters were placed, whenever possible, via the right common femoral vein (n = 33) to avoid inadvertent dislodgment of central thrombus during internal jugular venous insertion. For femoral insertion of the SVC filter, a jugular insertion kit was used so that the filter was oriented with the apex toward the heart. Similarly, a femoral insertion kit was used for jugular insertion of the SVC filter (n = 8). The filters were placed immediately below the confluence of the brachiocephalic veins, whenever possible, to protect the azygous vein from associated thromboembolic complications (Figs 1, 2). In patients with simultaneous lower extremity DVT, an IVC filter was placed after SVC filter placement, if femoral access was used, and before SVC filter insertion, if jugular access was used.
|
|
|
All chest radiographs obtained after filter placement were reviewed to determine the presence of filter migration, dislodgment, or fracture. The last chest radiograph obtained was used as the end of follow-up for filter migration, dislodgment, and fracture complications. The follow-up chest radiographs were posteroanterior in 22 patients (54%), posteroanterior and lateral in five patients (12%), and bedside in 14 patients (34%). Patients were followed up for immediate complications of filter placement such as evidence of subsequent PE, SVC syndrome, or upper extremity venous gangrene (phlegmasia cerulea dolens). Ventilation-perfusion imaging, pulmonary angiography, and computed tomography (CT) of the chest were not performed unless clinically indicated. Follow-up was maintained until death or until January 1998 in 38 patients. Three patients were lost to follow-up. Survival rates were determined by means of a Kaplan-Meir analysis.
| RESULTS |
|---|
|
|
|---|
A central venous catheter was present at the site of thrombosis at the time of diagnosis (n = 22) or within the previous 2 weeks (n = 14) in a total of 36 patients (88%). Several patients had hypercoagulable states secondary to heparin-associated antiplatelet antibodies (n = 7) or protein S deficiency (n = 2).
PE occurred prior to SVC filter insertion or systemic anticoagulation in eight patients. Filters were inserted because of complications related to anticoagulation (n = 23), contraindications to anticoagulation (n = 14), or for presurgical prophylaxis in the setting of substantial thromboembolic risk factors (n = 3). Heparin administration was reinstituted after filter placement in 10 patients (24%). Heparin administration initially had been stopped in these patients because of imminent surgery (n = 3), gastrointestinal bleeding (n = 4), intracranial hemorrhage (n = 2), or hemorrhagic pancreatitis (n = 1). One patient had filter placement in conjunction with systemic anticoagulation for propagating free-floating thrombus. Patients with a venous stenosis central to the upper extremity DVT did not have an SVC filter placed. Of the 16 patients (39%) with simultaneous lower extremity DVT, 15 underwent placement of an IVC filter. One patient with lower extremity DVT did not have an IVC filter placed for reasons that were not clear.
No patients had immediate or delayed complications that were directly related to the filter insertion.
At a median follow-up of 12 weeks (range, 1 day to 221 weeks) there was no evidence of filter migration, fracture, or dislodgment in any patient. Although 23 patients had subsequent placement of central catheters (n = 21) or Swan-Ganz catheters (n = 2), there was no filter dislodgment as a result.
There was no clinical evidence of SVC occlusion, venous gangrene, or exacerbation of upper extremity symptoms in any patient at a median follow-up of 15 weeks.
PE occurred in one patient 44 months after SVC filter insertion. The source of embolism in this patient was thought to be from the lower extremity, because the patient had acute left lower extremity DVT at duplex US. This patient did not have an IVC filter in situ. There was no upper extremity DVT or SVC thrombus evident at duplex US or magnetic resonance (MR) venography. No other patients in the study group had evidence of PE. In no patient was PE the cause of death. One autopsy was performed in the patient group. This patient had severe necrotizing pancreatitis, bronchopneumonia, and no evidence of PE. All other patients had a documented cause of death other than PE. Survival rates according to the Kaplan-Meir analysis were as follows: 1 week, 88%; 1 month, 73%; 6 months, 48%; and 12 months, 41% (Fig 3).
|
| DISCUSSION |
|---|
|
|
|---|
Prevention of PE by Means of Filter Placement
Langham et al (7) reported experimental placement of a Greenfield filter into the SVC of 11 dogs. Thrombus, harvested from phenolized segments of the IVC, was then embolized into the SVC filters. At 3-month follow-up, caval patency was maintained in all animals. There was no evidence of PE at autopsy in any dog.
Reports of SVC filter insertion in patients, although sparse and with short follow-up, have shown promising results. The only previous series (8) of which we are aware consisted of six patients who had PE due to upper extremity DVT despite anticoagulation. None of these patients had further PE 414 months after filter insertion.
There have been five case reports (913) of filter insertion into the SVC. In each case, the indication was PE due to upper extremity DVT despite anticoagulation. None of the patients had evidence of PE at follow-up of 3 months to 1 year.
One might expect the frequency of subsequent PE, despite the mechanical caval interruption, to be similar for SVC and IVC filters from upper and lower extremity sources, respectively. In patients with an IVC filter in situ, the frequency of recurrent, symptomatic PE is 2.4%4.0% (1523), with fatal PE occurring in less than 0.3% of cases (24). Direct comparisons cannot be made between the frequency of PE after IVC filter insertion and the frequency of PE after SVC filter insertion because of the much larger sample sizes in the IVC filter studies. Nevertheless, in our experience percutaneous filter insertion into the SVC was 100% effective in preventing subsequent symptomatic PE due to upper extremity DVT, and, in our study, the median and maximum follow-up was longer than those in previously reported studies. It is likely that reinstitution of anticoagulation in 10 patients added protection against PE in the study group.
In common with our study, the majority of published data about follow-up of IVC filter placement has been based on clinical criteria such as death or symptomatic PE. But these criteria are not optimal, because PE may be asymptomatic (25,26). Objective follow-up in all patients after filter placement, whether they are symptomatic for PE or not, is the ideal for future studies of filter effectiveness, regardless of filter location, and could be obtained by means of lung scintigraphy or pulmonary CT angiography.
Predisposing Factors for Upper Extremity DVT
The overall incidence of upper extremity DVT has increased from 1%2% of all DVT cases in the 1960s (27) to 4% of all DVT cases in 1988 (4). It is likely that this increased incidence is, in part, secondary to the increasing use of central venous access catheters in patient care. Aburahma et al (28) showed that the incidence of upper extremity venous thrombosis was secondary to central indwelling catheters in 17% of cases between 1982 and 1985 as compared with 47% of cases between 1986 and 1990. Furthermore, it has been reported (29) that secondary upper extremity DVT, particularly catheter-related thrombosis, results in PE more frequently than it does after primary upper extremity DVT.
Central venous access and Swan-Ganz catheters are likely to have contributed substantially to the development of upper extremity DVT in our patient group. Underlying malignancy, congestive cardiac failure, heparin-associated antiplatelet antibodies, and, in two patients, protein S deficiencies also predisposed patients to upper extremity DVT in this study.
Apart from treating the underlying predisposition to upper extremity DVT, the mainstay of treatment is rest, elevation of the affected arm, and anticoagulation. Thrombus propagation and the postphlebitic complications of DVT are unaffected by filter placement, therefore anticoagulation should be instituted or reinstituted whenever possible, as it was in 27% of our study group. Local fibrinolytic therapy is an effective method for restoring patency, particularly in patients with acute debilitating upper extremity DVT (30). Surgery is reserved for patients with an underlying cause for the thrombosis, such as thoracic outlet compression (30); also, thrombectomy occasionally is required (9).
Technique and Potential Complications
The technique of percutaneous filter insertion is of great importance to prevent complications and to ensure that the filter is positioned for optimum effectiveness. Initial documentation is required of thrombus extent, underlying venous stenoses, and SVC patency. Patients with upper extremity DVT and a central venous stenosis at our institution are generally regarded as being at a reduced risk for PE as compared with patients without central stenosis. The reasoning behind this is that the stenosis may prevent central migration of thrombus from the upper extremity, thereby acting as an "autofilter." We therefore do not routinely place SVC filters in such patients, however, to our knowledge, there are no published data yet to support this rationale.
Misplacement of an SVC filter is theoretically more likely than with an IVC filter because of the relatively short target area for deployment. Misplacements have been reported with IVC filter insertions, with one report (31) of misplacement into the right atrium. Although no SVC filter placements resulted in complications in our study, Langham et al (7) reported that in one of 11 dogs the limb of a misplaced Greenfield filter protruded into the brachiocephalic vein and resulted in caval perforation, which was detected at autopsy.
Symptomatic SVC occlusion did not occur in any patient in our study group. Nevertheless, SVC occlusion is one of the most dangerous complications that could occur after SVC filter placement. IVC occlusion has been reported to occur after filter insertion in less than 5% of cases (24), with asymptomatic IVC occlusion also reported after filter insertion (32). Occlusion of the SVC after SVC filter insertion has been described in two patients (12,13) with adenocarcinoma of the lung. The exact filter location was not reported in either case. It has been suggested, on the basis of these two reports, that SVC filter placement should be avoided in patients with bronchogenic carcinoma. Our results did not support this suggestion, however, because four patients with bronchogenic carcinoma and three patients with mediastinal adenopathy successfully underwent filter placement without complication. A superior vena cavogram must be obtained before filter insertion to ensure no involvement of the SVC by the mediastinal mass prior to filter insertion. We place the SVC filter immediately inferior to the confluence of the brachiocephalic veins, whenever possible, to avoid azygous vein occlusion should filter thrombosis occur. In addition, the placement more distal to the right atrium may reduce the risk of intracardiac migration. Careful deployment and close follow-up is recommended for each patient group. It must be reaffirmed that for femoral insertion of an SVC filter, a jugular set must be used for correct filter orientation and, therefore, effectiveness. A femoral set should similarly be used for jugular insertion of an SVC filter.
Choice of Filter
The Greenfield filter was used in the experimental dog study of Langham et al (7), as well as in 10 of the 11 previously described (812) human patients. The Vena Tech was deployed in the remaining case report (13).
In our study, no difference among the four filters used was detectable at follow-up. We used the Greenfield filter in the majority of cases primarily because of its longer history of safety in the IVC and the relatively small caval footprint. The Simon nitinol filter has the advantage of the smaller 9-F introducer, which can be deployed more easily through tortuous anatomy. Although we did not experience any complications in our one case, the Bird's Nest filter is probably the least favorable for use in the SVC because of its relatively long length (7 cm). If the patient is likely to require subsequent MR imaging examinations of the central chest veins, then use of the titanium Greenfield filter is optimal, because it produces the least artifact.
Placement of Central Catheters after Filter Insertion
A high percentage (56%) of our patients had central catheters placed after SVC filter insertion. In over half of these cases, the catheter was positioned through the filter. No complications resulted. In similar fashion, two patients in a previous study (8) underwent Swan-Ganz catheter placement after SVC filter placement without complication. Although central catheter placement after filter insertion appears to be safe, straight guide wires should be used at all times (33), because J guide wires could potentially hook onto and dislodge the filter. We believe that fluoroscopy should be used during insertion of central venous catheters if the catheter is to be placed adjacent to or advanced through the filter.
Appropriateness of SVC Filter Insertion
The survival rates for our patient group were low (eg, 48% at 6 months). This reflects the high frequency of malignancy and severe cardiac, respiratory, and intracranial diseases. The low survival rates were not attributable to upper extremity DVT or to PE. It is questionable whether SVC filter placement is an ethical use of limited resources in such patients. Practice guidelines are required, with further study, to determine if SVC filter placement is cost-effective, particularly in patients with metastatic disease.
Study Limitations
The low survival rates in our study resulted in a study limitation by detracting from our long-term outcome results as regards SVC filter insertion. A further limitation of our study was the lack of objective patient follow-up for PE, SVC perforation, and SVC occlusion. Objective follow-up of these potentially asymptomatic complications would be desirable in future studies of SVC filters. The exact indications for SVC filter placement also require further study. It is likely, however, that the indications will be akin to those for IVC filter placement. Finally, a filter designed specifically for the SVC would be desirable. This filter would require a shorter caval footprint and overall length than the current filters designed for IVC placement.
Conclusion
In our experience, percutaneous filter placement in the SVC is a safe and effective method for preventing symptomatic PE due to acute upper extremity DVT in patients in whom anticoagulation is contraindicated or ineffective. It is, however, recommended that anticoagulation be reinstituted after filter placement whenever possible. Placement of central catheters after SVC filter insertion is feasible but requires careful technique and fluoroscopic guidance.
| Footnotes |
|---|
Abbreviations: DVT = deep venous thrombosis IVC = inferior vena cava PE = pulmonary embolism SVC = superior vena cava
Author contributions: Guarantor of integrity of entire study, L.D.S.; study concepts and design, L.D.S., M.G.G.; definition of intellectual content, L.D.S., M.G.G.; literature research, L.D.S., M.G.G., H.M.M.; clinical studies, L.D.S., M.G.G., C.T.M., M.A.G., B.L.D.; data acquisition, L.D.S., C.T.M., H.M.M.; data analysis, L.D.S., M.G.G.; statistical analysis, L.D.S.; manuscript preparation, L.D.S.; manuscript editing, L.D.S., M.G.G., B.L.D.; manuscript review, L.D.S., M.G.G., H.M.M., C.T.M., M.A.G., B.L.D.
Received February 27, 1998;
revision requested March 30, 1998; revision received June 22, 1998;
accepted August 14, 1998.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. Kearon, S. R. Kahn, G. Agnelli, S. Goldhaber, G. E. Raskob, and A. J. Comerota Antithrombotic Therapy for Venous Thromboembolic Disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition) Chest, June 1, 2008; 133(6_suppl): 454S - 545S. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kadiev and P. Diaz A 47-Year-Old Man With Upper Extremity Pain and Edema and Severe Hypoxemia Chest, June 1, 2007; 131(6): 1982 - 1985. [Full Text] [PDF] |
||||
![]() |
H. R. Buller, G. Agnelli, R. D. Hull, T. M. Hyers, M. H. Prins, and G. E. Raskob Antithrombotic Therapy for Venous Thromboembolic Disease: The Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy Chest, September 1, 2004; 126(3_suppl): 401S - 428S. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. D. Stein, F. Kayali, and R. E. Olson Twenty-one-Year Trends in the Use of Inferior Vena Cava Filters Arch Intern Med, July 26, 2004; 164(14): 1541 - 1545. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. W. Rich The Management of Venous Thromboembolic Disease in Older Adults J. Gerontol. A Biol. Sci. Med. Sci., January 1, 2004; 59(1): M34 - 41. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Verso and G. Agnelli Venous Thromboembolism Associated With Long-Term Use of Central Venous Catheters in Cancer Patients J. Clin. Oncol., October 1, 2003; 21(19): 3665 - 3675. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. V. Joffe and S. Z. Goldhaber Upper-Extremity Deep Vein Thrombosis Circulation, October 1, 2002; 106(14): 1874 - 1880. [Full Text] [PDF] |
||||
![]() |
M. A Sheikh, A. P Topoulos, and S. R Deitcher Isolated internal jugular vein thrombosis: risk factors and natural history Vascular Medicine, August 1, 2002; 7(3): 177 - 179. [Abstract] [PDF] |
||||
![]() |
T. M. Hyers, G. Agnelli, R. D. Hull, T. A. Morris, M. Samama, V. Tapson, and J. G. Weg Antithrombotic Therapy for Venous Thromboembolic Disease Chest, January 1, 2001; 119(1_suppl): 176S - 193S. [Full Text] [PDF] |
||||
![]() |
Superior Vena Cava Filters for Upper-Extremity DVT Journal Watch (General), January 22, 1999; 1999(122): 8 - 8. [Full Text] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |