The D-dimer test is used commonly in diagnostic strategies to reduce the need for ultrasonography in patients suspected of having deep venous thrombosis. We studied several clinical and laboratory variables that might limit the accuracy of a semiquantitative D-dimer test.

In this retrospective cohort study, 704 outpatients suspected of having deep venous thrombosis underwent a semiquantitative D-dimer test and ultrasonography. The performance of the D-dimer test was calculated in patients using anticoagulants (n =61), patients with previous thrombosis (n =127), and patients with malignancy (n =47), including 39 patients with more than one of these characteristics. The 508 remaining patients were considered to be the reference group.

A total of 254 patients (36%) had evidence of deep venous thrombosis. The D-dimer test had a sensitivity of 99% (174/176; 95% confidence interval [CI]: 96% to 100%) and a negative predictive value of 98% (98/100; 95% CI: 93% to 100%) in the reference group. The sensitivity of the D-dimer test in patients using oral anticoagulants was 75% (6/8; 95% CI: 35% to 97%; P =0.01 compared with the reference group). Test sensitivity was 96% (51/53; 95% CI: 87% to 100%) in patients with previous thrombosis, and 100% (29/29; 95% CI: 88% to 100%) in patients with cancer. However, 553 (79%) of all patients, including 43 of the cancer patients (91%), had an abnormal D-dimer test.

The semiquantitative D-dimer test in this study has a high sensitivity and negative predictive value in the exclusion of deep venous thrombosis, except perhaps among patients using oral anticoagulants. D-dimer tests in patients with cancer and in patients over 70 years old may not be worthwhile, because the tests are usually positive.