Session I - Pelvic Trauma

Prospective Comparison of Contrast Enhanced Computed Tomography vs Magnetic Resonance Venography in the Detection of Occult Deep Pelvic Vein Thrombosis in Patients with Pelvic and Acetabular Fractures

Michael D. Stover, MD; Steven J. Morgan, MD; Michael J. Bosse, MD; Stephen H. Sims, MD; Brian J. Howard, MD; Daniel Stackhouse, MD; Matthew J. Weresh, MD; James F. Kellam, MD, Carolinas Medical Center Charlotte, NC

Purpose: To determine if the addition of intravenous contrast enhancement to routine pre-operative computed tomography scans (CT) is as effective as magnetic resonance venography (MRV) in detecting deep venous thrombosis within the pelvic venous system in patients with fractures of the acetabulum and pelvis, and to determine the false positive rate of each method by confirmatory selective pelvic venography.

Introduction: Deep venous thrombosis (DVT) is a common occurrence following traumatic injury. Risk factors for the development of DVT include injuries to the pelvis and acetabulum. More troublesome is the high rate of pulmonary embolism (PE) in these patients (2-12%). Effective screening of the pelvic venous system is a problem. To address this, Montgomery & Helfet, et al. introduced surveillance of the pelvic venous system in patients with acetabular fractures by magnetic resonance venography. Using MRV, they reported a DVT incidence of 22% in the pelvic veins. Ascending venography, however, could confirm only one pelvic vein thrombus. The limitations of MRV include limited availability, contraindications to magnetic resonance imaging, the inconvenience of an additional test for patients, and cost.

Patients with fractures of the acetabulum or pelvis routinely undergo CT evaluation of the bony injury. Contrast enhanced CT (CT venography) for the detection of DVT has been described, but its efficacy within the venous system is unknown. If CT venography is proven reliable for detection of DVT in the pelvic vasculature, it may be utilized as a cost-effective screening instrument for pelvic DVT with minimal intervention, and additional investigation with MRV can be avoided.

Materials and Methods: Patients between the ages of 16 and 69 with displaced pelvic and acetabular fractures admitted to the investigating institution, except for those with isolated posterior wall fractures, were potentially eligible for inclusion in the study. Subjects with known history of thromboembolic disease, previous vena caval filter placement, pelvic embolization coil placement, renal insufficiency, and routine MRI contraindications were excluded prospectively. Fractures were classified using the OTA classification system. DVT prophylaxis was administered at the time of admission based on a standard protocol. CT scans enhanced with intravenous contrast and MRV studies were performed concurrently within 24 hours of each other and usually within 48 hours of planned surgery. The CT venogram and MRV were evaluated independently by separate experienced radiologists blinded to the other study and its interpretation. If a thrombus was identified on either study, a selective pelvic venogram was performed to confirm the presence or absence of a thrombus.

Results: Seventy-four patients underwent surgery for a pelvic or acetabular fracture during the 14-month enrollment period. Thirty patients (19 males and 11 females) who met the inclusion criteria and consented to participate in the study were enrolled. The study group consisted of 10 pelvic ring injuries (OTA 8-61B, 2-61C) and 22 acetabular fractures (OTA 4-62A, 11-62B, 7-62C). The mean age of patients was 35 years (range 17-68). The mean ISS score was 15 (range 4-43). Four patients initially presented in shock, and 13 patients received a blood transfusion prior to imaging. Associated injuries considered to increase the risk of DVT included 1 spinal cord injury and 9 patients with at least one fracture of the lower extremity. Smoking history was significant in 17 patients, and obesity was a factor in 5 patients. Twenty-nine patients were immobilized at the time of presentation, 1 patient was ambulatory on crutches. The mean time between injury and study was 7 days (range 1-10). A DVT was identified by CT or MRV in 5 of the 30 patients (16.6%). A DVT was detected by CT scan in 2 patients (6.6%) and by MRV in 4 patients (13.3%). Of the 5 patients, only one patient was determined to have a DVT on both MRV and CT, but at different locations. Confirmatory selective pelvic venograms were performed in all 5 subjects. One venogram confirmed the presence of a DVT (3.3% occurrence rate). The one case of venogram-confirmed DVT was identified on CT venography only. The false positive rate for CT venography was 50%. The false positive rate for MRV was 100%. No cases of PE occurred postoperatively.

Discussion: The presence of an occult DVT in the pelvic vasculature may be more likely to result in PE with potential mortality. The ideal surveillance instrument for identification of DVT in the pelvic vasculature should be sensitive, specific, and non-invasive. It should allow evaluation of bilateral venous structures, be limited in time, and cost-effective. Most importantly, a screening tool should be validated against a known gold standard, in this case, pelvic venography. MRV is an attractive option for evaluating this area of the body in a non-invasive fashion. It is extremely sensitive but has diminished specificity, especially in areas of turbulent flow found in the pelvic vasculature. Montgomery & Helfet, et al. realized that the statistical accuracy of MRV for thrombus in the pelvic vasculature could not be definitively ascertained without direct cannulation through the femoral vein and selective venography. They relied upon its reported accuracy rate in the proximal femoral system and on previously published data to identify MRV as their "gold standard". Findings of MRV resulted in a change in management in 22% of patients in their initial series.

The patients in this study had significant risk factors for DVT and are comparable to those evaluated with MRV in previous studies. The incidence of pelvic DVT identified by MRV in this group (13.3%) is similar to that in the series reported by Montgomery and Helfet (22%). The absence of clot on confirmatory pelvic venogram in this study, however, suggests a much lower incidence of pelvic DVT and high false positive rate for DVT identified by MRV. Use of CT scan in this study was more effective than MRV in identifying the only confirmed case of pelvic DVT with a lower false positive rate. Contrast enhanced CT scanning is an attractive surveillance tool. It can evaluate bilateral venous structures, be obtained simultaneously with the routine pre-operative CT scan, decrease costs of additional study, and eliminate additional patient transfers and discomfort. It does carry the attendant risks of intravenous contrast injection.

Conclusion: In utilizing either MRV or CT venography as the screening tool for pelvic DVT, this study identifies the necessity of confirmatory selective pelvic venography prior to vena cava filter placement. CT venography was more effective than MRV in the identification of occult pelvic DVT.