“Filter or no filter” during percutaneous endovenous intervention for deep vein thrombosis

Mohsen Sharifi

By Mohsen Sharifi

Percutaneous endovenous intervention (PEVI) is increasingly being used in the treatment of massive lower extremity deep vein thrombosis. It utilises a variety of treatment modalities including catheter-directed thrombolysis, various thrombectomy devices, balloon venoplasty and stenting. There is a growing body of evidence that PEVI leads to a reduction in post-thrombotic syndrome and recurrent venous thromboembolism and a higher rate of venous patency.

Instrumentation in extensive amounts of acute thrombus, however, is potentially associated with iatrogenic pulmonary embolism. There are conflicting data as to the true incidence and clinical outcome and significance of patients with iatrogenic pulmonary embolism. Since the approaches used in PEVI are different, the incidence of iatrogenic pulmonary embolism would be variable. It stands to reason that bulkier catheters, more frequent device exchanges and runs and prolonged duration of manipulation inside the thrombus would lead to a higher rate of iatrogenic pulmonary embolism. For this reason the incidence of iatrogenic pulmonary embolism has been reported as low as 0–1% when using catheter-directed thrombolysis alone vs. 33–45% when bulkier catheters are used. Consequently it has been a dilemma especially for the novice operator whether or not to prophylactically place an inferior vena cava filter before PEVI.

Until recently, there were no randomised trials specifically addressing the role of inferior vena cava filter implantation before PEVI for deep vein thrombosis. The recently published FILTER-PEVI (Filter implantation to lower thrombo-embolic risk in PEVI) trial by our group, randomised 141 patients undergoing PEVI for acute massive deep venous thrombosis to receive a prophylactic inferior vena cava filter (filter group=70 patients) to no filter (control group=71 patients). The baseline clinical characteristics, anticoagulation and PEVI regimen (Table 1) were similar between the two groups. Follow-up was obtained at 15±2 months.

Table 1. Utilised interventional approaches


  Filter Control Group
  n= 70 n=71
Trellis 34 36
Angiojet 8 9
Thrombolytic therapy via infusion catheter 32 35
Balloon venoplasty 56 54
Stent 18 16

Patients developing symptoms suggestive of pulmonary embolism during or within 24 hours of PEVI underwent objective testing for pulmonary embolism. Fourteen patients of the filter group and 22 of the control group underwent such testing in whom new pulmonary embolism was detected in one and eight patients, respectively (1.4% vs. 11.3% of the total population, p=0.048).

There was, therefore, an eightfold increase in symptomatic iatrogenic pulmonary embolism in those not receiving a filter. However, no increased mortality was noted in patients without a filter. Profound hypotension and bradycardia requiring resuscitory measures occurred in three patients (4.2%) of the control group during PEVI as a result of iatrogenic pulmonary embolism. These were transient events, and the patients recovered without any sequelae. The 11.3% rate of iatrogenic pulmonary embolism in the control was in keeping with the 8.6% of venographic evidence of entrapped thrombus by the filter in the filter group. No pulmonary embolism developed in patients undergoing catheter-directed thrombolysis alone.

Development of venous thromboembolism after hospital discharge was similar between the two groups which was limited to deep venous thrombosis in two patients of each group. No patient developed pulmonary embolism at follow-up. This study identified five predictors of iatrogenic pulmonary embolism (Table 2).

Table 2. Predictors of iatrogenic pulmonary embolism during PEVI 

Predictor p value
Pulmonary embolism on admission p<0.001
Inflammatory form of deep vein thrombosis* p=0.044
Two or more adjacent venous segments involved p=0.004
Venous diameter of ≥7mm p=0.042
Preserved venous anatomy p=0.045

P-values were calculated using a Z-test for proportions, two-tailed

*Marked tenderness, erythema, induration and oedema with rapid onset and progression

The FILTER-PEVI trial was the first randomised trial to evaluate the role of inferior vena cava filters in the reduction of iatrogenic pulmonary embolism during PEVI for deep venous thrombosis. Based on the findings of this trial, several conclusions can be made: 1) Inferior vena cava filters do reduce the incidence of iatrogenic pulmonary embolism by a factor of 8; 2) Development of iatrogenic pulmonary embolism leads to significant symptoms in only 4.2% of the patients who do not receive a filter without affecting short- or mid-term mortality and recurrent venous thromboembolism; 3) The noted symptoms are transient and uniformly occur at the time of the procedure and not later; 4) Those receiving catheter-directed thrombolysis alone do not require a filter; 5) Nor would patients with chronically venosclerotic segments wherein the lesser thrombus load, narrowed vessel diameter, and uneven endoluminal surface (with consequent higher resistance to free flow of debris) reduces the embolisation risk. 6) A selective approach in placement of filters may be exercised in patients with predictors of iatrogenic pulmonary embolism who will have additional intervention to catheter-directed thrombolysis alone; 7) If an operator chooses not to place a filter in patients with predictors of iatrogenic pulmonary embolism (excluding catheter-directed thrombolysis alone cases), they should be prepared for a 4.2% chance of dealing with some form of “excitement” in the angiography suite.

Mohsen Sharifi is director, Arizona Cardiovascular Consultants & Vein Clinic, and adjunct professor, AT Still University, Mesa, USA.