Diagnostic characteristics of lower limb venous compression ultrasonography in suspected pulmonary embolism: a meta-analysis

Background: Diagnosis of pulmonary embolism (PE) is commonly based on D-dimer measurement and computed tomography (CT) angiography. Lower limb vein compression ultrasonography (CUS) for diagnosing deep vein thrombosis may be of interest in patients with suspected PE. Objectives: We aimed to summarize the data on the diagnostic characteristics of CUS in suspected PE patients. Patients/Methods: We conducted a literature review by using PUBMED and EMBASE and included 15 prospective studies in which CUS was performed in consecutive patients with suspected PE. Results: Of the 6991 included patients, 2001 (30%) had pulmonary embolism. Eight of the 15 studies included only outpatients, two included hospitalized patients and five involved both in-and outpatients. In 13 studies, only proximal CUS was performed. Two studies analyzed the added value of distal CUS including the calf veins (whole-leg CUS). Pooled estimate of proximal CUS sensitivity was 41% (95% confidence interval [CI], 36-46%) with strong heterogeneity (I square, 79%). Specificity of proximal CUS was 96% (95% CI, 94-98%). The overall positive likelihood ratio for proximal CUS was 11.9 (95% CI, 7.1-19.8), whereas the overall negative likelihood ratio was 0.6 (95% CI, 0.5-0.7). The sensitivity of whole-leg CUS was 79% (95% CI, 24-98%) and specificity was 84% (95% CI, 76-90%). Conclusions: Proximal CUS has low sensitivity and cannot be used to rule out PE. Nevertheless, its high specificity allows confirming PE, which may be useful in patients with contraindications to CT angiography. Whole-leg CUS has a higher sensitivity but low specificity for PE and can therefore not be recommended.