Correspondence
Vena Caval Filters for the Prevention of Pulmonary Embolism
N Engl J Med 1998; 339:46-48July 2, 1998
- Article
To the Editor:
In their report of the efficacy and safety of vena caval filters in patients with proximal deep-vein thrombosis, Decousus et al. (Feb. 12 issue)1 conclude that the initial benefit of vena caval filters for the prevention of pulmonary embolism was counterbalanced by an excess of recurrent deep-vein thrombosis, without any difference in mortality.
In their study, the filters did reduce the rate of mortality from pulmonary embolism within 12 days after randomization. Four of five patients in the no-filter group died of pulmonary embolism, as compared with none of five patients in the filter group. Although the overall mortality rates were similar, patients in the filter group died of bleeding, myocardial infarction, respiratory failure, or renal failure. None of these causes of death are known adverse effects of the filters.2 Therefore, mortality from pulmonary embolism is probably a better indicator of the efficacy of the filters than overall mortality.
Did the use of filters reduce the risk of symptomatic pulmonary embolism after two years? On the basis of the data presented, the answer is probably yes. Although the P value was 0.16, it cannot be concluded that there was no treatment effect. The estimated cumulative incidence of symptomatic pulmonary embolism within two years was 3.4 percent in the filter group and 6.3 percent in the no-filter group. With these incidence rates, 200 patients in each group, and an alpha level of 0.05, the statistical power is only 0.27. In other words, there is a 73 percent risk of erroneously drawing a negative conclusion. Failure to detect a treatment effect was mainly due to small samples.3,4 Indeed, the filters reduced the risk of symptomatic pulmonary embolism by 46 percent.
Patients in the filter group had an increased risk of recurrent deep-vein thrombosis, but a decreased risk of symptomatic pulmonary embolism. One possible explanation for this finding is that the filters were effective in preventing symptomatic pulmonary embolism.
Deep-vein thrombosis is of much less clinical concern than pulmonary embolism, especially symptomatic pulmonary embolism. The benefit of filters for the prevention of pulmonary embolism could not be “counterbalanced” by an excess of recurrent deep-vein thrombosis. In view of the reduced incidence of and mortality from pulmonary embolism with the use of vena caval filters, as shown in this study, it is prudent to continue our current practice with respect to this issue.
4 ReferencesWen-Min Chuu, M.D., M.P.H.
Washington Hospital Center, Washington, DC 20010Nae-Yuh Wang, M.S.
Johns Hopkins University, Baltimore, MD 21205David Perry, M.D.
Washington Hospital Center, Washington, DC 200101
Decousus H ,Leizorovicz A ,Parent F , et al. A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. N Engl J Med 1998;338:409-415
Full Text | Web of Science | Medline2
Becker DM ,Philbrick JT ,Selby JB . Inferior vena cava filters: indications, safety, effectiveness. Arch Intern Med 1992;152:1985-1994
CrossRef | Web of Science | Medline3
Freiman JA ,Chalmers TC ,Smith H Jr ,Kuebler RR . The importance of beta, the type II error and sample size in the design and interpretation of the randomized control trial: survey of 71 “negative“ trials. N Engl J Med 1978;299:690-694
Full Text | Web of Science | Medline4
Moher D ,Dulberg CS ,Wells GA . Statistical power, sample size, and their reporting in randomized controlled trials. JAMA 1994;272:122-124
CrossRef | Web of Science | Medline
To the Editor:
Decousus et al. state that they undertook their study “to assess the efficacy and safety of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis.” However, Decousus et al. studied patients who did not have the accepted indications for vena caval filters, and they did not include in their study a group of patients with acute deep venous or pulmonary thromboembolic disease who received vena caval filters without anticoagulant therapy. The authors' conclusions are not supported by their research because of the absence of a treatment group consisting of patients with acute disease who received only vena caval filters.
The current standard of care for patients presenting with pulmonary or deep venous thromboembolic disease is anticoagulant therapy or placement of a vena caval filter. Patients who undergo filter placement, most of whom do not receive anticoagulant therapy, have less than a 5 percent incidence of pulmonary embolism,1-3 which is similar to the incidence in patients receiving anticoagulant therapy.
Decousus et al. report that the incidence of pulmonary embolism within 12 days after randomization was 4.8 percent in the patients receiving anticoagulant therapy without vena caval filters and 1.1 percent in those receiving anticoagulant therapy with filters (P=0.03). The incidence of recurrent pulmonary embolism in the first 12 days in patients presenting with pulmonary embolism was reduced from 8.6 percent to 1.1 percent by the use of vena caval filters in addition to anticoagulation. The authors report no clinically significant complications of filter placement. Of the 10 patients who died within 12 days after beginning treatment, 2 receiving anticoagulant agents died from hemorrhagic complications, and 4 in the no-filter group died of pulmonary embolism. It would have been interesting if the investigators had included a group of patients who had presented with pulmonary embolism alone, and it would also be interesting to know how many patients were excluded because they were not considered to be at high risk for pulmonary embolism, despite their having proximal deep-vein thrombosis.
In summary, this study does not support the conclusion that vena caval filters do not prevent pulmonary embolism in the absence of anticoagulant drugs. It does suggest higher morbidity associated with anticoagulant therapy than with the placement of vena caval filters and confirms the higher early incidence of fatal pulmonary embolism in patients receiving anticoagulant drugs without vena caval filters.
3 ReferencesTimothy P. Murphy, M.D.
Brown University School of Medicine, Providence, RI 02906Scott O. Trerotola, M.D.
Indiana University School of Medicine, Indianapolis, IN 46202Robert L. Vogelzang, M.D.
Northwestern University School of Medicine, Chicago, IL 606111
Greenfield LJ ,Michna BA . Twelve-year clinical experience with the Greenfield vena caval filter. Surgery 1988;104:706-712
Web of Science | Medline2
Murphy TP ,Dorfman GS ,Yedlicka JW , et al. LGM vena cava filter: objective evaluation of early results. J Vasc Interv Radiol 1991;2:107-115
CrossRef | Medline3
Ferris EJ ,McCowan TC ,Carver DK ,McFarland DR . Percutaneous inferior vena caval filters: follow-up of seven designs in 320 patients. Radiology 1993;188:851-856
Web of Science | Medline
To the Editor:
Decousus et al. are to be commended for their randomized study of vena caval filters in patients with proximal deep-vein thrombosis, which showed a statistically significant difference in the incidence of pulmonary embolism between the filter and no-filter groups (only 1.1 percent in the filter group and 4.8 percent in the no-filter group, P=0.03). However, they chose to criticize filters because of the possibility that they are associated with an increased incidence of recurrent deep-vein thrombosis within two years. Such an association is unlikely unless there is distal venous hypertension when the filter traps a massive embolus. This has occurred in only 3 percent of a population of patients with Greenfield filters whom we have followed for more than 20 years.1
The authors incorrectly assume that all filters are the same. Although all filters trap emboli, their long-term patency varies. Objective testing of Greenfield filters has established a 95 percent patency rate, whereas the alternative devices have 15 to 20 percent lower rates of patency. The authors should indicate whether they noted any differences.
It is interesting that the incidence of major bleeding and related deaths has been glossed over. With a total of 39 major bleeding events (10 percent) and 9 of 83 deaths due to bleeding (11 percent), this complication was considerably more serious than recurrent deep-vein thrombosis.
From the standpoint of methodology, the possibility of selection bias is strong, since more than 40 centers enrolled patients with the use of subjective criteria. These patients differ from those in whom filters are usually placed, limiting the generalizability of the findings.
1 ReferencesLazar J. Greenfield, M.D.
Mary C. Proctor, M.S.
University of Michigan, Ann Arbor, MI 48109Author/Editor Response
The authors reply:
To the Editor: The reference treatment for acute venous thromboembolism is anticoagulant therapy, so the only two well-accepted indications for filters in this setting are contraindications to or failure of anticoagulation.1 In the United States, however, 20 to 50 percent of vena caval filters are placed for prophylactic purposes.1,2 The aim of our study was to evaluate a potential additive benefit of prophylactic filters in patients with proximal deep-vein thrombosis who were treated with anticoagulant agents. Since up to 3.4 percent of cases of pulmonary embolism occur in patients treated with both filters and anticoagulant agents,3 it does not seem ethical to evaluate a group of patients with deep-vein thrombosis treated with filters alone, as suggested by Murphy et al., when there is no contraindication to anticoagulant therapy.
In our study, the filters used at a center were always the same; a center did not use several types of filters. Neither randomization nor stratification took into account the type of filter. Therefore, the filter's effect is confounded with the center's effect. Moreover, the data presented by Greenfield and Proctor4 are debatable, since data were missing for more than 20 percent of the patients (140 of 642) in their retrospective study. We believe that Greenfield and Proctor have minimized the clinical importance of recurrent deep-vein thrombosis. Of the patients with recurrent deep-vein thrombosis in the filter group, 10 percent had pulmonary embolism and 43 percent had thrombosis of the filter.3 Moreover, an excess recurrence of deep-vein thrombosis may increase the incidence of the postphlebitic syndrome.
As Chuu et al. suggest, a nonsignificant effect on mortality due to pulmonary embolism could be a good indicator of the efficacy of filters if other data on efficacy and safety showed the same favorable trend. However, this was not the case. Moreover, of the four fatal cases of pulmonary embolism in patients without filters, only one occurred in the group receiving the low-molecular-weight heparin.3 Thus, the difference in fatal pulmonary embolism between the groups may not be attributable only to an effect of the filters, a finding in agreement with other data in the literature.5 However, we agree with Chuu et al. that filters are associated with a trend toward a reduced risk of symptomatic pulmonary embolism at two years. Only longer follow-up, currently being carried out, can confirm such a trend.
5 ReferencesHervé Decousus, M.D.
Patrick Mismetti, M.D.
Bernard Tardy, M.D.
Centre Hospitalier Universitaire Saint-Etienne Bellevue, 42055 Saint-Etienne CEDEX 2, FranceFor the Prévention du Risque d'Embolie Pulmonaire par Interruption Cave Study Group
1
Becker DM ,Philbrick JT ,Selby JB . Inferior vena cava filters: indications, safety, effectiveness. Arch Intern Med 1992;52:1985-1994
CrossRef | Web of Science2
Greenfield LJ ,Proctor MC . Endovascular methods for caval interruption. Semin Vasc Surg 1997;10:310-314
Medline3
Decousus H ,Leizorovicz A ,Parent F , et al. A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. N Engl J Med 1998;338:409-415
Full Text | Web of Science | Medline4
Greenfield LJ ,Proctor MC . Twenty-year clinical experience with the Greenfield filter. Cardiovasc Surg 1995;3:199-205
CrossRef | Medline5
Leizorovicz A . Comparison of the efficacy and safety of low molecular weight heparins and unfractionated heparin in the initial treatment of deep venous thrombosis: an updated meta-analysis. Drugs 1996;52:Suppl 7:30-37
CrossRef | Web of Science | Medline
- Citing Articles (6)
Citing Articles
1
Miguel Angel De Gregorio, Pablo Gamboa, Diana L. Bonilla, Maitane Sanchez, Maria T. Higuera, Jokin Medrano, Antonio Mainar, Fernando Lostalé, Alicia Laborda. (2006) Retrieval of Günther Tulip Optional Vena Cava Filters 30 Days after Implantation: A Prospective Clinical Study. Journal of Vascular and Interventional Radiology 17:11, 1781-1789
CrossRef2
Joseph Cherian, Elie Gertner. (2005) Recurrent Pulmonary Embolism Despite Inferior Vena Cava Filter Placement in Patients With the Antiphospholipid Syndrome. JCR: Journal of Clinical Rheumatology 11:1, 56-58
CrossRef3
Christopher S. Morris, Frederick B. Rogers, Kenneth E. Najarian, Anant D. Bhave, Steven R. Shackford. (2004) Current Trends in Vena Caval Filtration with the Introduction of a Retrievable Filter at a Level I Trauma Center. The Journal of Trauma: Injury, Infection, and Critical Care 57:1, 32-36
CrossRef4
Thomas B. Kinney. (2003) Update on Inferior Vena Cava Filters. Journal of Vascular and Interventional Radiology 14:4, 425-440
CrossRef5
Thomas W Wakefield. (2000) Treatment options for venous thrombosis. Journal of Vascular Surgery 31:3, 613-620
CrossRef6
M SIMON. (1999) Vena Cava Filters: Prevalent Misconceptions
. Journal of Vascular and Interventional Radiology 10:8, 1021-1024
CrossRef
