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


Compression ultrasound was not useful in symptom-free venous thrombosis

ACP J Club. 1994 Nov-Dec;121:78. doi:10.7326/ACPJC-1994-121-3-078

Source Citation

Jongbloets LM, Lensing AW, Koopman MM, Büller HR, ten Cate JW. Limitations of compression ultrasound for the detection of symptomless postoperative deep vein thrombosis. Lancet. 1994 May 7;343:1142-4.



To compare compression ultrasound with contrast venography for detecting deep venous thrombosis (DVT) in symptom-free patients after craniotomy.


Blinded comparison of compression ultrasound and contrast venography findings.


Tertiary care center in the Netherlands.


115 patients (mean age 61 y, 55% women) who had had a craniotomy and had no contraindication to anticoagulant treatment or contrast venography.

Description of Test and Diagnostic Standard

Compression ultrasound was done using a 7.5-MHz linear-array sonographic scanner to examine the common femoral, superficial femoral, and popliteal veins and the upper halves of the peroneal and posterior tibial veins. The criterion for DVT was incomplete vein compression by the ultrasound transducer as seen in transverse view. Bilateral ascending contrast venography was done immediately after ultrasonography. Criteria for DVT were an intraluminal filling defect present on at least 2 different projections or a venous segment that did not fill despite repeated injections of contrast material.

Main outcome measures

Sensitivity, specificity, and positive predictive value of compression ultrasound compared with the diagnostic standard of ascending contrast venography for detecting proximal DVT and calf DVT.

Main results

Proximal DVT was detected in 13 patients by venography and in 5 of those patients by ultrasound. Ultrasound showed false-positive results for proximal DVT in 4 patients. Of 91 patients in whom ultrasound did not indicate proximal DVT, 71 had evaluable findings for calf vein assessment. Calf DVT was detected in 16 patients by venography and in 8 of those patients by ultrasound. Ultrasound showed false-positive results for calf DVT in 15 patients. Overall, of 26 patients with venography-confirmed DVT, ultrasound detected 13. Of 32 positive results with ultrasound, 19 were false positives (positive predictive value 41%, CI 24% to 60%). Sensitivities, specificities, and likelihood ratios are shown in the Table.


Compression ultrasound was not useful in detecting deep venous thrombosis in symptom-free patients after craniotomy.

Source of funding: Dutch Thrombosis Foundation.

For article reprint: Dr. A.W. Lensing, Department of Neurosurgery and Neurology, Academic Medical Centre, F4-237, Melbergdreef 9, 1105 AZ Amsterdam, the Netherlands. FAX 31-412-662537.

Table. Test characteristics of compression ultrasonography for detecting proximal deep venous thrombosis (DVT) and calf DVT*

Outcomes Sensitivity (95% CI) Specificity (CI) +LR -LR
Proximal DVT 38% (8 to 69) 95% (88 to 99) 7.6 0.65
Calf DVT 50% (25 to 75) 73% (59 to 84) 1.85 0.68
All DVT 50% (29 to 71) 74% (63 to 84)† 1.92 0.68

*LRs defined in Glossary and calculated from data in article.
†Calculated from data in article.


Compression ultrasonography is highly sensitive and specific for the detection of proximal DVT in symptomatic patients with clinically suspected DVT (1). Recent studies (2), however, have shown that the sensitivity of compression ultrasonography is approximately 60% for proximal DVT in postoperative orthopedic patients. The likely explanation for the lower sensitivity of ultrasound in postoperative orthopedic patients is the tendency for their thrombi to be smaller, less occlusive, and, therefore, less likely to be detected than thrombi in symptomatic patients after surgery (2).

The study of Jongbloets and colleagues in patients having craniotomy is sound methodologically, and its results are consistent with those reported in contemporary studies evaluating compression ultrasonography in postoperative orthopedic patients. Although only a few patients (n = 13) had proximal DVT, the upper limit of the 95% CI on the observed sensitivity is 69%, a result that is inconsistent with a high sensitivity and is unacceptable clinically. They also showed that compression ultrasonography has low positive predictive value for any DVT and low sensitivity for calf DVT. The high proportion of small nonocclusive thrombi and the high sensitivity of ultrasound for large occlusive thrombi but not for smaller nonocclusive thrombi are noteworthy. Therefore, clinicians should be aware that patients having craniotomy commonly develop DVT (incidence of 29% in this study), that a normal ultrasound scan at discharge does not reliably exclude proximal or calf DVT, and that an abnormal ultrasound scan should be confirmed by venography. In the future, technologic advances may improve the accuracy of ultrasound scans for detecting DVT in postoperative patients.

Jeffrey S. Ginsberg, MD
Chedoke-McMaster HospitalsHamilton, Ontario, Canada


1. Lensing AW, Prandoni P, Brandjes D, et al. Detection of deep-vein thrombosis by real-time B-mode ultrasonography. N Engl J Med. 1989;320:342-5.

2. Ginsberg JS, Caco CC, Brill-Edwards PA, et al. Venous thrombosis in patients who have undergone major hip or knee surgery: detection with compression US and impedance plethysmography. Radiology. 1991;181:651-4.