Value of ultrasound imaging of the lower limbs for the diagnosis of deep vein thrombosis in an emergency department

Objectives: 1) To describe the characteristics of patients with these symptoms. 2) To

assess agreement between compression ultrasonography performed in the ED and

Doppler ultrasound imaging in radiology. 3) To assess the efficacy of a diagnostic

algorithm that includes ultrasound imaging, clinical characteristics, and laboratory

findings.

Methods: Descriptive study of consecutive patients attending the ED complaining of

symptoms suggestive of DVT. We applied a diagnostic algorithm requiring physical

examination, D-dimer assessment, pretest probability (Wells score), and ED compression

ultrasonography.

Results: Thirty-seven patients were included; 48.65% were men. Pretest probability

according to the Wells score was low for 9 patients (24.3%; 95% confidence interval

[CI], 9.1%-39.5%), moderate for 10 (27.0%; 95% CI, 11.4%-42.7%), and high for 18

(48.6%; 95% CI, 31.2%-66.1%). D-dimer level was not significantly associated with the

findings of compression ultrasonography. The ê index for agreement between

compression ultrasonography and the gold standard (Doppler ultrasound in the

radiology department) was 1. Sensitivity was 100% (95% CI, 96.1%-100%), specificity

91.7% (95% CI, 71.9%-100%) , positive predictive value 92.9% (95% CI, 75.8%-

100%), and negative predictive value 100% (95% CI, 95.4%-100%).

Conclusions: The algorithm tested appears to be valid for evaluating patients who may

have DVT. There is a high level of agreement between compression ultrasonography of

the lower extremities in the ED and ultrasound findings reported by the radiology

department.