Is Prophylactic Vena Caval Filter Placement Beneficial in Acute Spinal Cord Injury?
Peter H.U. Lee, MD, Haisar Dao, MD, Ronald Gross, MD, Lisa Patterson, MD, Patrick C. Lee, MD
St. Elizabeth’s Medical Center, Boston, MA
PURPOSE OF STUDY
Acute spinal cord injury (SCI) patients are at high risk for venous thromboembolism (VTE) and fatal pulmonary embolism (PE) is a well-recognized risk. The use of prophylactic vena caval filters (VCF) in this population is controversial.
The Nationwide Inpatient Sample from the Healthcare Cost and Utilization Project was used to identify SCI patients between 2001 and 2005. Data were collected regarding the incidence of DVT and PE, and the use of VCF based on ICD-9 codes. Mortality rates were also determined. The number needed to treat (NNT) with prophylactic VCF to prevent a PE fatality was estimated assuming that 20% of DVTs result in PEs and 10% of PEs are fatal.
SUMMARY OF RESULTS
In this sample, there were 39.3 million patient discharges, of whom 19,549 were identified with SCI. There was a 2.5% incidence of DVT and 1.0% incidence of PE, with 96.5% of SCI patients had neither DVT nor PE. 1573 patients received VCF, of whom 11.3% had DVT, 4.3% had PE, while the remaining 84.4% had neither DVT nor PE and were considered to have had prophylactic placement of a VCF. From 2001 to 2005, the overall rate of VCF use in SCI pts had increased from 214 to 409. For patients with no DVT or PE, there was no difference in the in-hospital mortality rate with (7.4%) or without (6.8%) VCF placement (p=0.33; RR=1.11; 95% CI 0.91, 1.35). Based on a 5.6% and 1.8% incidence of DVT and PE, respectively, in patients who survived beyond 3 days, the NNT with prophylactic VCF to prevent a fatal PE is 910 patients.
The majority of VCF placed in SCI patients were for prophylaxis. However, prophylactic use of VCF does not appear to lower the overall mortality rate in these patients. Given the relatively low incidence to DVT and PE, and a large resultant NNT to prevent a single fatal PE, the potential benefits of routine prophylactic VCF placement may be limited and is not supported by these findings.