Deep venous thrombosis (DVT) of a lower extremity is a commonly encountered clinical problem that can lead to potentially fatal outcome if pulmonary embolism develops. D-dimers, end products of fibrin clot degradation, have been used as markers for the presence of venous clots. Wells and associates reported on a randomized trial that incorporated the use of a D-dimer assay to risk-stratify the patients who require ultrasonography for assessment of possible DVT.

The investigators recruited 1,285 patients from several academic health centers who presented with symptoms suspicious for DVT, of whom 1,096 (85 percent) consented to enrollment and had no clinical exclusion criteria (e.g., resolution of initial symptoms, suspected pulmonary embolism). Participants were first screened with a clinical prediction model for the likelihood of DVT (see accompanying table on page 2006). Patients with scores of two or more were considered likely to have DVT. Participants were randomized to ultrasonography alone ord-dimer testing followed by ultrasonography if needed. All patients randomized to ultrasonography alone had an initial imaging examination that was repeated one week later if the result of the first ultrasound examination was negative but the clinical model predicted that DVT was likely. Patients randomized to thed-dimer assay- only regimen had ultrasonography if the clinical model predicted an increased risk of DVT or if the assay was positive. All trial participants were then followed for three months to determine if unsuspected DVT developed. Complete follow-up was available in 1,082 patients (99 percent).

Overall, DVT occurred in 15.7 percent of patients. Of the 566 patients randomized to the use of the D-dimer assay who had a negative assay result (and negative ultrasound result if the clinical model predicted increased risk of DVT), a venous clot was later identified during follow-up in two patients (0.4 percent). The rate of subsequently detected DVT was 1.4 percent among those initially deemed negative for clot by ultrasonography alone. Incorporation of the D-dimer assay allowed physicians to arrive at a definitive diagnosis on the initial day of presentation in 82 percent of patients compared with 65 percent of those randomized to ultrasonography alone.

The authors conclude that if their clinical prediction model deems DVT unlikely and the D-dimer assay is negative for clot, the diagnosis of DVT can be excluded safely without the need for ultrasonography.

editor's note: of the 566 patients randomized to use of the d-dimer test, negative assay results allowed 299 patients (53 percent) to be definitively diagnosed without the need for additional ultrasonography. an accompanying editorial by bockenstedt¹ notes with caution that a number of different d-dimer assays are available, each with various technical limits. it is not entirely clear that the high negative predictive value of the d-dimer assays employed in this research trial can be extrapolated to use in the general medical community.—b.z.