Emergency Department compression ultrasound to diagnose proximal deep vein thrombosis

Abstract

Emergency Department (ED) patients with suspected deep vein thrombosis (DVT) require an objective vascular study such as ultrasound (US) to confirm the diagnosis prior to treatment or disposition. A simple compression US test of the common femoral vein and popliteal vein reliably detects proximal DVT in symptomatic patients. Application of compression US in the ED by Emergency Physicians (EPs) has been tested in a single previous study. We evaluated the ability of ED compression US, performed by EPs, to diagnose proximal DVT as compared to duplex US performed in a vascular laboratory. A prospective, observational study was conducted on a convenience sample of patients presenting to an ED with lower extremity symptoms and signs suggestive of DVT. Patients with a history of DVT in the symptomatic extremity were excluded. Final diagnosis of DVT was made by color-flow duplex US performed in a vascular laboratory. ED compression US was performed by one of six EP sonographers. In compression US, DVT was diagnosed by the inability to compress the common femoral vein or popliteal vein. The examination was considered indeterminate if the veins could not be clearly identified or compressibility was equivocal. For statistical analysis, an indeterminate examination was considered positive. In those cases where ED compression US was discordant with duplex US, and not indeterminate, we retrospectively reviewed the US findings. There were 76 patients who completed the study, and 18 patients (24%) were diagnosed with DVT by duplex US, among whom ED compression US was positive in 14, negative in 2, and indeterminate in 2. Among 58 patients diagnosed without DVT by duplex US, there were 4 false-positive ED compression US examinations and 10 indeterminate examinations. In all, ED compression US was indeterminate in 12 patients (15.8%). Compared to duplex US, ED compression US had a sensitivity of 88.9% (95% C.I. 65.3-98.6%) and specificity of 75.9% (62.8-86.1). Negative predictive value was 95.7% (85.2-99.5). Among ED patients with the clinical diagnosis of possible DVT, negative ED compression US greatly reduces the likelihood of DVT, such that discharge and outpatient follow-up can be considered. Because of limited specificity, positive results require confirmation, but may justify immediate treatment pending follow-up testing. Indeterminate results can be expected in a significant number of patients and mandate further testing prior to disposition.