Deep vein thrombosis: vena caval filters were not helpful in combination with anticoagulation.

Clinical bottom line (level 1b)

  1. Patients with proximal DVT who were at high risk of pulmonary embolism and had a vena caval filter together with anticoagulation, had fewer pulmonary embolisms in the first 12 days than those not given a filter (NNT = 29 at 12 days) , but had no clear difference over two years.
  2. Patients given a vena caval filter were more likely to have a recurrent DVT within two years (NNH = 13 at 2 years) .
Decousus et al: New England Journal of Medicine 1998; 338 (7): 409-415
Expires May 2003

The study

Unblinded concealed randomised trial without intention-to-treat
Setting: 44 centres, France

400 patients (aged range 60 to 94 years; mean 73, 50% male) acute proximal deep vein thrombosis, diagnosed by venogram, and considered to be at high risk for pulmonary embolism by their physician

Excluded if
  • <18 years old
  • contraindication to or failure of anticoagulation therapy
  • previous filter inserted
  • on curative anticoagulation >48 hours
  • indication for thrombolysis
  • short life expectancy
  • iodine allergy
  • hereditary thrombophilia
  • severe hepatic or renal failure
  • pregnant
  • patient unlikely to be compliant


    • Control Group: (n = 200, 200 analysed): no intervention
    Experimental Group: (n = 200, 200 analysed): inferior vena caval filter. Filters were inserted percutaneously using fluoroscopic control via femoral or jugular vein
    All patients received iv heparin (5000 units bolus followed by 500 mg/kg/day titred to aPTT 1.5 to 2.5) or s/c low molecular weight heparin (enoxaparin bd) by random allocation, followed by warfarin or acenocoumarol started on day four, titred to INR 2.0 to 3.0 for three months. All patients also wore stockings for three months. If unable to have warfarin, had s/c LMWH instead.
    100% followed for 12 months (reviewed after 3 and 12 months, and contacted by phone after 2 years)

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNT
    (95% CI)
    pulmonary embolism 12 days 9
    (4.5%)    		 2
    (1.0%)    		 78%
    (-2% to 95%)    		 3.50%
    (0.31% to 6.69%)    		 29
    (15 to 320)
    symptomatic pulmonary embolism 12 days 5
    (2.5%)    		 2
    (1.0%)    		 60%
    (-100% to 92%)    		 1.50%
    (-1.07% to 4.07%)    		 67
    (NNT = 94 to infinity;
    NNH = 25 to infinity)
    major bleed 12 days 6
    (3.0%)    		 9
    (4.5%)    		 -50%
    (-310% to 46%)    		 -1.50%
    (-5.22% to 2.22%)    		 -67
    (NNT = 19 to infinity;
    NNH = 45 to infinity)
    death 12 days 5
    (2.5%)    		 5
    (2.5%)    		 0%
    (-240% to 71%)    		 0.00%
    (-3.06% to 3.06%)    		 inf
    (NNT = 33 to infinity;
    NNH = 33 to infinity)
    symptomatic pulmonary embolism 2 years 12
    (6.0%)    		 6
    (3.0%)    		 50%
    (-31% to 81%)    		 3.00%
    (-1.05% to 7.05%)    		 33
    (NNT = 95 to infinity;
    NNH = 14 to infinity)
    recurrent DVT 2 years 21
    (10.5%)    		 37
    (18.5%)    		 -76%
    (-190% to -7%)    		 -8.00%
    (-14.9% to -1.14%)    		 -13
    (-87 to -7)
    major bleed 2 years 22
    (11%)    		 17
    (8.5%)    		 23%
    (-41% to 58%)    		 2.50%
    (-3.31% to 8.31%)    		 40
    (NNT = 30 to infinity;
    NNH = 12 to infinity)
    death 2 years 40
    (20%)    		 43
    (22%)    		 -7%
    (-58% to 27%)    		 -1.50%
    (-9.45% to 6.45%)    		 -67
    (NNT = 11 to infinity;
    NNH = 16 to infinity)

    Comments

    1. It is unclear how high risk patients were defined.
    2. A number of important clinical questions remain. Would patients have received only positive benefits if filters had been removed after a few weeks? How would patients with thrombophilia differ? Are filters an alternative to anticoagulation?

    Citation

    1. 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. New England Journal of Medicine 1998; 338 (7): 409-415
    Contributor: Chris Ball and Clare Wotton, May 2000
    Reviewer: Charles Fisher

    Clinical Question.
    Patient DVT
    Intervention or Exposure vena caval filter
    Comparison no intervention
    Outcome PE and death