Journal of Vascular Surgery
Volume 8, Issue 4 , Pages 460-464, October 1988

Percutaneous insertion of the Greenfield inferior vena cava filter: Experience with ninety-six patients☆

Presented at the Twelfth Annual Meeting of the Southern Association for Vascular Surgery, St. Thomas, Virgin Islands, Jan. 27-30, 1988.

Departments of Diagnostic Radiology (Drs. Pais, Tobin, and Austin) and Surgery (Dr. Queral), University of Maryland Medical System/Hospital. Baltimore, Md.

Article Outline

Abstract 

This article evaluates the ease, safety, and convenience of percutaneous Greenfield filter placement and compares percutaneous with surgical placement. Greenfield filters were inserted percutaneously into the inferior vena cava in 96 patients. Ninety filters were placed via the femoral route and 12 were placed from the right internal jugular vein. Six patients had two filters inserted. An inferior venacavogram was performed before filter insertion in all patients. Cavography provided vital information concerning diameter of the inferior vena cava, the level of the renal veins, and the presence and location of thrombus. Filter placement was accomplished in all patients in whom it was attempted. There were four minor complications and one periprocedural death. The incidence of documented femoral vein thrombosis that could be related to percutaneous placement via the femoral veins was 33%; however, none of these patients had permanent venous stasis sequelae. Percutaneous insertion of the Greenfield filter is a safe and convenient procedure and is superior to surgical placement in terms of time, logistics, and the accuracy of filter positioning. (J VASC SURG 1988;8:460-4.)

 

The Greenfield filter has been in use since 1973 and is the most widely used method of inferior vena cava (IVC) interruption in this country.1, 2 The usual method of inserting this device is by means of a cutdown of the right internal jugular vein.3 Percutaneous insertion of this filter via both the jugular and femoral approach was first described in 1984.4 Since then, several series have been reported in the radiologic literature.5, 6, 7 This article describes our experience with the percutaneous insertion of Greenfield filters in 96 patients. Safety, ease of insertion, and convenience are evaluated and compared with surgical insertion.

Back to Article Outline

Material and methods 

Between June 1986 and November 1987, Greenfield filters were percutaneously placed in 96 patients. There were 52 male and 44 female patients ranging in age from 18 to 80 years (mean 54 years). The patients' preexisting disorders included recent major surgery in 54 patients (52%), neurologic abnormalities in 40 patients (42%), major trauma in 32 patients (33%), malignancy in 32 patients (33%), cardiopulmonary disease in 16 patients (17%), and sepsis in five patients (5%). (The total is more than 96 patients and more than 100% because some patients are represented in two or more categories.) The indications for filter placement were contraindication to anticoagulation or a complication of anticoagulation in a patient with proven pulmonary embolus (42 patients [43%]), recurrent embolism while adequately anticoagulated (three patients [3%]), and prophylaxis (51 patients [53%]).

The method of filter insertion has been previously described.7 Cavography is done before all filter insertions. Heparin administration is discontinued six hours before the procedure and warfarin is stopped at least 48 hours before filter placement. A radiograph is obtained after insertion to document filter position. Orders given after the procedure include bed rest for 4 hours and observation of the insertion site at half-hour intervals for 4 hours. The entire procedure can usually be performed in approximately 30 minutes.

To assess the incidence of femoral vein thrombosis related to percutaneous placement via the femoral route, 24 patients were examined with either duplex scans (10 patients) or portable ultrasound (14 patients) if the patient could not be easily transported. The ultrasound examinations were performed by one radiologist (K. D. T.) with a 7.5 MHz transducer (model ATL MK 100, Advanced Technology Laboratories, Bothell, Wash.). The duplex scans were performed by vascular technologists in the vascular laboratory. The examination was performed with a 7.5 MHz transducer (Ultrasonics 750 SD Duplex, Ultrasonics, Yonkers, N.Y.). Patients were examined 3 to 34 days after filter insertion (mean 14 days). Six patients with abnormal results of noninvasive studies had phlebography. Two additional patients had noninvasive studies consistent with femoral vein thrombosis, but the referring clinicians refused phlebography.

Patient follow-up was based on information obtained from the referring physicians and from medical records. The follow-up period ranged from 1 to 16 months (mean 8 months).

Back to Article Outline

Results 

During the course of this study 98 patients were referred to the Department of Radiology for percutaneous filter placement. Filter insertion was not attempted in two patients. One patient had a caval diameter greater than 30 mm and a filter was not inserted because of the danger of filter migration. This patient was seen early in our series; currently we place bilateral iliac filters in patients with a caval diameter greater than 30 mm. The second patient had caval thrombosis extending almost to the level of the right atrium, thereby precluding filter placement. Percutaneous filter placement was technically successful in the remaining patients. One hundred two filters were inserted in 96 patients, 90 via the femoral route and 12 from the right internal jugular approach. The indications for jugular insertion were bilateral iliac vein or caval thrombosis (10 patients) or failure to negotiate the pelvis with the filter carrier (two patients). Two filters were inserted in each of six patients. In the first patient, who had deep venous thrombosis (DVT) of the upper extremity and pulmonary emboli, a filter was placed in the superior vena cava from the femoral approach.8 Another filter was also placed in the IVC because a pelvic source of emboli could not be excluded. Four patients had IVC diameters ranging from 33 to 42 mm and a filter was placed in each common iliac vein. The sixth patient had a high junction of the iliac veins and a filter was inadvertently released in the right iliac vein. This was immediately recognized and a second filter was introduced from the left iliac vein and appropriately positioned in the IVC. There were no adverse sequelae.

There were four minor complications; all were due to technical errors and occurred early in the series. As described above, one filter was misplaced in an iliac vein, resulting in a misplacement incidence of 1%. In a second patient, the 24F sheath used to introduce the filter carrier was pushed under the skin during the attempt to insert the carrier through the sheath. Initial attempts to retrieve the sheath were not successful. Bleeding was profuse and not controllable with manual compression; it was controlled by passing an 8 mm balloon catheter over the wire into the sheath under fluoroscopic control and inflating the balloon. A skin incision was made along the course of the catheter until the sheath became visible, and it was retrieved with the use of a hemostat. The damaged sheath was replaced, and the procedure was completed without further incident. This complication is no longer possible because the current sheath used for the procedure has a protective sleeve, making it impossible to advance under the skin.

In a third patient, a pneumothorax developed the day after filter insertion from the jugular approach, and a chest tube was required. The final minor complication was tilting of a single filter. Retrospective analysis of the cavogram demonstrated a small venous tributary at the level of the foot that extended beyond the margin of the IVC; therefore this complication could have been avoided.

One death occurred that was related to filter placement. A 71-year-old woman was admitted to our Shock Trauma unit after a motor vehicle accident with the clinical diagnosis of stroke, although the CT scan of the head done at admission did not demonstrate any evidence of acute infarction. During the course of her hospitalization clinical evidence developed of lower extremity DVT. Cavography showed thrombus extending above the level of the renal veins. A filter was placed from the right internal jugular vein. Approximately 2 to 5 minutes after the carrier was removed, the patient became unresponsive, tachycardic, and hypotensive, requiring intubation. A head CT scan performed approximately 18 hours after filter insertion was unchanged from the admission study. The patient died 18 hours later. The autopsy failed to demonstrate any evidence of cerebral infarction, pulmonary thromboembolism, or air embolism. The cause of death was attributed to congestive heart failure and atherosclerotic cardiovascular disease. The exact cause of this death remains unclear. There were two cases of documented IVC thrombosis after filter insertion. In both cases anticoagulation therapy was begun when their clinical status allowed, and neither patient had permanent venous stasis sequelae.

Eight of the 24 patients evaluated for femoral vein thrombosis were found to have thrombus formation at the site of filter insertion (33%); however, only three of the patients were symptomatic. Two of the symptomatic patients later died of causes related to their underlying diseases; the remaining symptomatic patient had lower extremity swelling, which resolved after anticoagulation therapy. On follow-up no patient had venous stasis sequelae.

Excluding the procedure-related death, seven other patients died within 2 weeks of filter placement. Three died of their underlying cardiopulmonary disease, two of adult respiratory distress syndrome, and two of gastrointestinal hemorrhage. Nine deaths occurred between 1 and 3 months after filter placement; none was related to filter placement or recurrent emboli. Twenty patients were lost to follow-up. The remaining 59 patients had no evidence of insertion site complications or recurrent pulmonary emboli.

The inferior venacavogram (IVCG) provided information that affected filter placement in 26% of patients. An IVC larger than 30 mm in diameter was found in five patients. Six patients were found to have an IVC that varied markedly in diameter along its length, requiring exact filter placement. One of these patients had an IVC markedly narrowed above the level of the renal veins (Fig. 1).

Filter insertion from the jugular approach would not have been possible, and a filter was successfully placed from the left femoral vein.

In six patients a large thrombus was present at the IVC bifurcation (Fig. 2), precluding filter placement from either iliac vein.

In these patients insertion was done from a jugular approach. An additional six patients were found to have either unilateral or bilateral iliac vein thrombosis. In four patients the renal veins were noted to be below the usual L-1 to L-2 level; one patient had a pelvic kidney with renal veins at the L-3 to L-4 level.

Back to Article Outline

Discussion 

Our minor complication rate was 4% with a misplacement incidence of 1%. This compares favorably with the incidence of operative complications reported in multiple surgical series.9, 10, 11, 12, 13, 14, 15 The incidence of misplaced surgically inserted filters is not clear, but several studies indicate misplacement rates of 1.5% to 12.5%.9, 10, 11 The superior radiographic imaging that is available in an angiographic suite greatly aids precise filter placement. Image distortion is most likely to occur with portable fluoroscopic equipment used by technically inexperienced personnel. Attempting to count the vertebral bodies on fluoroscopy, even when done by a radiologist using good equipment, can be confusing and misleading.

We have found that by placing a radiopaque ruler on the table at the time of the IVCG, these problems can be eliminated. The filter can be placed by reference to the numbers on the ruler, with assured accuracy. Of course, this depends on the patient not being moved between the performance of the IVCG and filter placement. Precise filter positioning was required in 15% of our patients because of caval thrombus, low renal veins, or variation in caval diameter.

The exact cause of the one death in our series remains unclear. The autopsy demonstrated no evidence of cerebral air embolism or massive pulmonary embolism from dislodging thrombus in the IVC. Although there has been no other reported death associated with Greenfield filter insertion,2 the more widespread use of this device increases the probability of random individual fatalities.

Our incidence of IVC thrombosis (2%) is in accord with statistics in the surgical literature.10, 11, 12, 13, 16

In the surgical literature the preferred site of filter introduction has been the right internal jugular vein,3 for two main reasons. First, there is essentially a straight line between the site of insertion and the location at which the filter is placed. This is in contrast to the curve that must be negotiated as the carrier traverses the iliac vein. Second, because of this straight line, the filter is deposited in the center of the cava; thus there is supposedly less tendency for the filter to be tilted in relation to the caval axis. This has not proved to be true in our experience, in which only 1 of 81 caval filters inserted via the femoral approach was tilted. Although the interventional radiologist should be equally adept at placing the filter from either the right internal jugular or the femoral vein, the femoral approach is generally preferred for the following reasons: there is no risk of air embolism or pneumothorax; the jugular route may result in accidental puncture of the carotid artery; the same puncture site can be used for pulmonary angiography, cavography and filter insertion; in the alert patient, working from the groin is less threatening than working from the neck; the radiologist is more experienced with the femoral approach.

However, the incidence of femoral vein thrombosis has emerged as a significant problem after femoral vein insertion. The incidence of this complication in several percutaneous series in which patients were monitored clinically varies from zero to 12%.5, 6, 17 A more recent study documented with phlebography a 41% incidence of thrombosis related to filter placement.18 Our findings are similar. Although the percentage of patients having femoral vein thrombosis is high, the number who are symptomatic is low. None of the patients in our series had venous stasis sequelac. The planned introduction of a 12F delivery system for the Greenfield filter (personal communication, Margaret Belmont, Meditech, Inc., Watertown, Mass.,) may ameliorate or eliminate this one significant problem with the femoral vein route for percutaneous filter insertion.

An IVCG is a prerequisite to filter insertion. Vital information concerning the safety of filter insertion, route of insertion, and proper positioning of the filter is obtained from this study. In the surgical literature there has been little emphasis on the importance of an IVCG. As recently as 1986 a case report of a filter migrating to the heart, in which an IVCG was apparently not done, suggests that this procedure is still not being universally performed before filter placement.19 The fact that cavography provided essential information in 26% of our patients confirms its importance. In our series 5 of 97 patients (5%) had a caval diameter greater than 30 mm; however, the base of the filter measures only 30 mm. Our findings suggest that the risk of filter migration is substantial if filters are placed without the benefit of cavography.

The reported incidence of migration is limited to scattered case reports.19, 20, 21 This probably results from the anteroposterior diameter of the IVC being less than the transverse diameter, thereby preventing migration. However, even without migration, a filter placed in such a cava might be suboptimal. For example, if all of the legs were not engaged, there could be an increased incidence of tilting, as well as the possibility of emboli passing between the filter and the caval wall.

A recent report describes manual expansion of the struts of a Greenfield filter before placement in a patient with a large IVC.9 However, further discussion in the same report suggests that this should not be done because it alters the geometry of the filter and therefore decreases its efficacy. We altered several filters by inserting a thumb into the apex of the cone and expanding the struts until the diameter at the base of the filter measured 40 mm. When the filter was then inserted into the carrier and released, the diameter of the base had returned to 30 mm. This practice is dangerous, not because it changes the geometry of the filter, but because it does not in fact alter filter size. In patients with caval diameters greater than 30 mm, we recommend placement of bilateral iliac filters. In an occasional patient this may not be possible because of thrombosis of one or both iliac veins. In this event, it would be necessary to consider an alternate method of caval interruption.

In most surgical series the percentage of filters placed prophylactically is between 10% and 24%.12, 13, 14, 16 An exception is the report by Cantelmo et al.22 in which 47 of 70 patients (67%) had filters placed for prophylaxis. However, this entire group consisted of cancer patients, who have both a high incidence of hypercoagulability and problems with anticoagulation. In our series 53% of filters were placed for prophylaxis, which is considerably higher than any other series with a general patient population. This was due in part to the large number of cancer, neurosurgical, and trauma patients in our population who have contraindications to anticoagulation. Our criteria for prophylaxis are in accord with Jones et al.2 and Greenfield.12

In summary, the morbidity and mortality of percutaneous and surgical placement of the Greenfield filter are comparable. Percutaneous insertion is more convenient and the filter can be inserted at the same time that diagnostic studies such as pulmonary angiography and phlebography are performed. The superior imaging available in the angiographic suite allows for more accurate filter positioning. Cavography, which is vital for proper filter placement, is a routine part of the procedure. The incidence of femoral vein thrombosis needs further study. If it proves to be a significant problem, then the preferred route of percutaneous insertion will have to be reevaluated.

Back to Article Outline

References 

  1. Messmer JM, Greenfield LJ. Greenfield caval filters: long-term radiographic follow-up study. Radiology. 1985;156:613–618
  2. Jones TK, Barnes RW, Greenfield LJ. Greenfield vena caval filter: rationale and current indications. Ann Thorac Surg. 1986;42(Suppl):S48–S55
  3. Mansour M, Chang AE, Sindelar WF. Interruption of the inferior vena cava for the prevention of recurrent pulmonary embolism. Am Surg. 1985;51:375–380
  4. Tadavarthy SM, Casteneda-Zuniga W, Salomonowitz E, et al.  Kimray-Greenfield vena cava filter: percutaneous introduction. Radiology. 1984;151:525–526
  5. Denny DF, Cronan JJ, Dorfman GS, Esplin C. Percutaneous Kimray-Greenfield filter placement by femoral vein puncture. AJR. 1985;145:827–829
  6. Rose BS, Simon DC, Hess ML, Van Aman ME. Percutaneous transfemoral placement of the Kimray-Greenfield vena cava filter. Radiology. 1987;165:373–376
  7. Pais SO, Mirvis SE, DeOrchis DF. Percutaneous insertion of the Kimray-Greenfield filter: technical considerations and problems. Radiology. 1987;165:377–381
  8. Pais SO, DeOrchis DF, Mirvis SE. Superior vena caval placement of a Kimray-Greenfield filter. Radiology. 1987;165:385–386
  9. Carabasi RA, Moritz MJ, Jarrell BE. Complications encountered with the use of the Greenfield filter. Am J Surg. 1987;154:163–168
  10. Cimochowski GE, Evans RH, Zarins CK, Lu CT, DeMeester TR. Greenfield filter versus Mobin-Uddin umbrella, the continuing quest for the ideal method of vena caval interruption. J Thorac Cardiovasc Surg. 1980;79:358–365
  11. Greenfield LJ, Peyton R, Crute S, Barnes R. Greenfield vena caval filter experience: late results in 156 patients. Arch Surg. 1981;116:1451–1456
  12. Greenfield LJ. Current indications for and results of Greenfield filter placement. J Vasc Surg. 1984;1:502–504
  13. Gomez GA, Cutler BS, Wheeler HB. Transvenous interruption of the inferior vena cava. Surgery. 1983;93:612–619
  14. Golueke PJ, Garrett WV, Thompson JE, Smith BL, Talkington CM. Interruption of the vena cava by means of the Greenfield filter: expanding the indications. Surgery. 1988;103:111–117
  15. Scurr JH, Jarrett PE, Wastell C. The treatment of recurrent pulmonary embolism: experience with the Kimray-Greenfield vena cava filter. Ann R Coll Surg Engl. 1983;65:233–234
  16. Wingerd M, Bernhard VM, Maddison F, Towne JB. Comparison of caval filters in the management of venous thromboembolism. Arch Surg. 1978;113:1264–1271
  17. Weinstein JK, Ramchandani P, Soulen RL, Desai SA. Complications of the percutaneous femoral approach for Kimray-Greenfield vena caval filter placement. In: Presented at the Annual Meeting of the Radiological Society of North America, Chicago. November 1987;
  18. Kantor A, Glanz S, Gordon DH, Sclafani SJA. Percutaneous insertion of the Kimray-Greenfield filter: incidence of femoral vein thrombosis. AJR. 1987;149:1065–1066
  19. Friedell ML, Goldenkranz RJ, Parsonnet V, et al.  Migration of a Greenfield filter to the pulmonary artery: a case report. J Vasc Surg. 1986;3:929–931
  20. Moore R, Dagher FJ, Tavares S, Attar S. Migration of Kimray-Greenfield umbrella to the heart. South Med J. 1983;76:946–947
  21. Casteneda F, Herrera M, Cragg AH, et al.  Migration of a Kimray-Greenfield filter to the right ventricle. Radiology. 1983;149:690
  22. Cantelmo NL, Menzoian JO, LoGerfo FW, Fasulo G, Mozden PJ. Clinical experience with vena caval filters in high-risk cancer patients. Cancer. 1982;50:341–344

☆ Reprint requests: S. Osher Pais, MD, Dept. of Diagnostic Radiology, University of Maryland Medical System/Hospital, 22 South Greene St., Baltimore, MD 21201.

PII: 0741-5214(88)90111-5

doi:10.1067/mva.1988.avs0080460

Journal of Vascular Surgery
Volume 8, Issue 4 , Pages 460-464, October 1988

Article Tools

* Image Usage & Resolution