Routine prophylactic vena cava filtration is not indicated after acute spinal cord injury

Acute spinal cord injury (SCI) is a devastating problem, with over 10,000 new cases annually. Pulmonary embolism (PE) is a well-recognized risk in SCI patients, although no clear recommendations for prophylaxis exist. We therefore evaluated whether routine placement of prophylactic inferior vena cava filters is indicated in SCI patients.

The trauma registry of a regional trauma center was used to identify patients sustaining acute SCI resulting in tetraplegia or paraplegia after blunt or penetrating trauma for a 5-year period beginning in January 1995. Patients were analyzed for demographics, mechanism of injury, Injury Severity Score, associated long bone or pelvic fracture, severe closed-head injury, type of deep venous thrombosis (DVT) prophylaxis, level of SCI, and incidence of DVT and PE. DVT prophylaxis was performed in all patients with sequential compression devices (SCDs) when extremity fracture status permitted. Data are expressed as mean +/- SD and analyzed using Fisher's exact test.

There were 8,269 admissions during the study period, with an overall incidence of DVT and PE of 11.8% and 0.9%, respectively. There were 111 (1.3%) patients who sustained SCI, with an incidence of DVT and PE of 9.0% and 1.8%, respectively, and no deaths. Of these 111 patients, 41.4% were paraplegics and 58.6% were tetraplegics, and 17.1% of patients had severe closed-head injury. Mean hospital length of stay was 23 +/- 20 days for SCI patients. Surveillance duplex ultrasound was performed an average of 2.3 +/- 2.1 times during each hospitalization. Mean Injury Severity Score was 30.0 +/- 12.2. The incidence of DVT and PE in those patients with SCDs alone was 7.1% and 2.3%; for SCDs plus subcutaneous heparin, the incidence was 11.1% and 2.8%; and for SCDs plus low-molecular-weight heparin, the incidence was 7.4% and 0%, respectively, with no statistical difference between groups. The incidence of DVT in SCI patients with long bone fractures was 37.5%, which was significantly greater than the total SCI population (p < 0.02).

The incidence of DVT and PE in SCI patients was similar to that of the overall trauma population when appropriate DVT prophylaxis was used. Subgroup analysis demonstrated that SCI associated with long bone fracture significantly increases the incidence of DVT. On the basis of the low incidence of PE in the present study, routine placement of prophylactic caval filters does not appear warranted in all SCI patients. However, SCI patients with long bone fractures, patients with DVT formation despite prophylactic anticoagulation, or patients with contraindications to anticoagulation may be appropriate candidates for prophylactic caval filtration.