Current issues of ACP Journal Club are published in Annals of Internal Medicine


Compression ultrasonography detected proximal deep venous thrombosis but not isolated iliac and calf deep venous thrombosis

ACP J Club. 1992 Mar-April;116:54. doi:10.7326/ACPJC-1992-116-2-054

Source Citation

Pedersen OM, Aslaksen A, Vik-Mo H, Bassoe AM. Compression ultrasonography in hospitalized patients with suspected deep venous thrombosis. Arch Intern Med. 1991 Nov;151:2217-20.



To assess the accuracy of compression ultrasonography in detecting deep venous thrombosis in hospitalized patients.


Independent evaluations of compression ultrasonograms and contrast venograms were compared.


Norwegian medical center.


Consecutive, hospitalized patients suspected of having deep venous thrombosis were referred for contrast venography. 91% of patients were referred from the medical department and 9% were referred from the surgical department. 215 patients had successful venograms (3 patients had both legs investigated). The 112 women and 103 men had a mean age of 64 years (range 16 to 91 years).

Description of test and diagnostic standard

Ultrasound examinations included transverse and longitudinal scanning of iliac, femoral, and popliteal veins. Criteria for deep venous thrombosis were intraluminal soft-tissue echoes and lack of total luminal compressibility. The proximal extent of the thrombus was defined as the point at which the thrombosed vein became compressible. The calf veins were not investigated.

The standard for diagnosis was contrast venography. Views of the veins of the calf, knee, thigh, and pelvis were obtained. Acute deep venous thrombosis was defined as the presence of an intraluminal filling defect, abrupt termination of a column of contrast medium, and diversion of blood flow.

Main outcome measure

Diagnosis of deep venous thrombosis. Comparisons were made separately for proximal and distal veins.

Main results

Deep venous thrombosis was detected by venography in 142 of 215 patients (66%). Of these thrombi, 29 (20%) did not extend beyond the calf vein and, therefore, were not examined by ultrasonography. 113 patients were diagnosed with proximal deep venous thrombosis. 101 of these were identified by ultrasonography (sensitivity 89%, 95% CI 83% to 95%). Venography identified no thrombosis in 73 patients; ultrasonography identified thrombi in 2 of these (specificity 97%, CI 93% to 100%). Ultrasonography identified thrombosis better in 103 patients with thrombi above the knee than in 10 patients with distal popliteal thrombi (sensitivity 93% vs 50%, P = 0.006). Sensitivity for pelvic-vein thrombosis (34 patients) was 71%.


Compression ultrasonography accurately identified proximal deep venous thrombosis in patients hospitalized with suspected thrombi. This method was, however, insensitive to the occurrence of isolated iliac and calf deep venous thrombosis.

Source of funding: Not stated.

Address for article reprint: Dr. O.M. Pedersen, Department of Clinical Physiology, Haukeland Hospital, University of Bergen, N-5021 Bergen, Norway.


Compression ultrasonography (CU) is becoming the most widely used noninvasive technique for the diagnosis of venous thrombosis. The popliteal to common femoral veins, but not the calf veins, can be examined by this technique, which is sensitive and specific to symptomatic proximal vein thrombosis in ambulant patients (1). However, CU is relatively insensitive to asymptomatic proximal thrombi, which tend to be much smaller than symptomatic thrombi (2). This well-designed study by Pedersen and colleagues was done on hospitalized patients, who had a 66% prevalence of venous thrombosis, about twice the prevalence seen in outpatients (1). CU was found to be clinically useful because its high specificity (97%) allowed management decisions to be made in 111 of the 215 patients. However, CU was only moderately sensitive to proximal vein thrombosis (89%) and was insensitive to calf vein thrombosis, a potentially important limitation because approximately 20% of calf vein thrombi extend proximally if left untreated. The lower sensitivity of CU in hospitalized patients than in outpatients (1) might be explained by technical difficulties or by small thrombi in the distal popliteal veins, which extended infrequently into the proximal veins.

Three different management approaches are currently used to manage symptomatic patients with negative CU. 1) Proceed as if venous thrombosis has been excluded (this is potentially dangerous because it leaves patients with venous thrombosis untreated); 2) repeat the test several times over the next week to detect extending thrombi that are not detected at the initial examination (this approach has been used successfully in outpatients); and 3) perform a venogram and base a management decision on the result. Either of the latter 2 approaches is acceptable.

Jack Hirsh, MD
Hamilton Civic HospitalsHamilton, Ontario, Canada


1. Lensing AW, Prandoni P, Brandjes D, et al. N Engl J Med. 1989;320:342-5.

2. Borris LC, Christiansen HM, Lassen MR, et al. Thromb Haemost. 1989;61:363-5.