In previous studies, enzyme-linked immunosorbent assays (ELISA) for plasma D-dimer analysis have demonstrated high sensitivity, suggesting their potential usefulness in excluding deep venous thrombosis (DVT). We evaluated the usefulness of a new D-dimer test (Liatest D-dimer) for suspected DVT in a prospective study of patients admitted to the hospital because of recent (not exceeding 1 week before admission) clinical signs. Contrast venography or compression ultrasonography or both were performed within 24 hours of admission. A new quantitative determination of D-dimer concentration using a suspension of microlatex particles coated with specific antibodies was tested. A standard plasma D-dimer ELISA measurement was also performed. Of 464 patients, 276 had a proven DVT (distal, 74; proximal, 202). For a cutoff level of 400 ng/mL, sensitivity of the Liatest method in the diagnosis of overall DVT was 94.6% (95% confidence interval, 92.0%-97.0%), and the specificity was 35% (95% confidence interval, 28%-42%). The sensitivity and negative predictive value were 98.5% and 95.6%, respectively, in the diagnosis of proximal DVT, but only 83.8% and 84.6%, respectively, in the diagnosis of distal DVT. This new rapid Liatest D-dimer assay seems to be highly sensitive and could replace the ELISA method in excluding patients with proximal DVT. Both methods provide lower sensitivity for distal DVT.