Objective To summarize the probability of pulmonary embolism （PE） induced by lower extremity deep venous thrombosis （DVT） and investigate the role of vena cava filter （VCF） in preventing from PE. Methods The clinical data of 1 058 patients with lower extremity DVT from January 2005 to January 2012 were analyzed retrospectively. Results The PE rate was 3.21% （34/1 058） and the death rate was 1.42% （15/1 058） in 1 058 patients with lower extremity DVT. The VCF was implanted in 171 of 1 058 patients. The VCFs of 151 patients were implanted from femoral vein， 20 patients were implanted from jugular vein. The PE rates were 3.61% （32/887） and 1.17% （2/171） and the death rates were 1.69% （15/887） and 0 （0/171） in patients without VCF and with VCF， respectively. Both of them occurred in the first ten days. PE could keep as long as 35 d. The PE rate and death rate in the patients without VCF were significantly higher than those in the patients with VCF （P＜0.01）. The PE rates and death rates in both lower extremities DVT were higher than those in patients with the right and left ones （P＜0.05）， which in the right lower extremity were higher than those in the left one （P＜0.05）. The PE rate and death rate in the patients with lower extremity DVT combined with vena cava thrombosis were significantly higher than those in the patients with central type （P＜0.05）， which in the central type were significantly higher than those in the peripheral type （P＜0.05）， there were no significant differences between peripheral type and mixed pattern. The follow-up time was from 1 month to 7 years with （39±19） months， the patency rate of VCF was 98.7%. There were no filter migration， declination， and failure of expansion. Conclusions VCF can prevent from PE effectively， but the indications must be controlled.