The (Absence of a) Role of Venous Duplex Scan in the Diagnosis of Pulmonary Embolism after Severe Trauma

Abstract

Venous duplex scan (VDS) has been used for interim bedside diagnosis of pulmonary embolism (PE) in severely injured patients deemed to be at risk if transported out of the intensive care unit. In combination with the level of clinical suspicion for PE, VDS helps select patients for temporary treatment until definitive diagnosis is made. We evaluate the sensitivity and specificity of VDS in critically injured patients with a high level of clinical suspicion for PE. We performed a prospective observational cohort study at the surgical intensive care unit of an academic level 1 trauma center. Patients were 59 critically injured patients suspected to have PE over a 30-month period. The level of clinical suspicion for PE was classified as low or high according to preset criteria. Interventions were VDS and a PE outcome test (conventional or computed tomographic pulmonary angiography). The sensitivity and specificity of VDS to detect PE in all patients and in patients with high level of clinical suspicion was calculated against the results of the outcome test. PE was diagnosed in 21 patients (35.5%). The sensitivity and specificity of VDS was 33 per cent and 89 per cent, respectively. Among the 28 patients who had a high level of clinical suspicion for PE, the sensitivity of VDS was 23 per cent and the specificity 93 per cent. In this latter population, 1 of the 4 (25%) positive VDS was of a patient without PE and 10 of the 24 (42%) negative VDS were of patients who had PE. VDS does not accurately predict PE in severely injured patients, even in the presence of a high level of clinical suspicion.