Summary. This prospective study was designed to assess the diagnostic sensitivity, specificity and negative predictive value of the NycoCard D-dimer plasma immunofiltration assay in patients with suspected deep vein thrombosis (DVT) confirmed by ultrasonography/venography. 84 medical patients were recruited: 43 patients (51%) had proven venous thrombosis, 33 by venography and 10 by ultrasonography. The sensitivity of NycoCard D-dimer in patients with DVT was 95 . 3%, the specificity was 22 . 0% and the negative predictive value was 81 . 8%. An algorithm including the NycoCard D-dimer test for the acute management of DVT is proposed. This would enable low-risk patients to be discharged early from hospital, without imaging or anticoagulant therapy.Keywords: deep vein thrombosis, D-dimer, diagnosis, laboratory tests.The lifetime risk of developing a DVT is approximately 2-5% (Davies & Tuddenham, 1987). Diagnosis and management of this condition therefore forms a substantial part of the acute workload in most hospitals.The clinical diagnosis of DVT is unreliable. The 'gold standard' for diagnosis is ascending radiographic venography, which has significant morbidity: local pain, allergic reactions to the contrast material and thrombosis induced by the procedure itself. In addition, the delay before a definitive result is known incurs extra cost if protective hospital admission and anticoagulation occur.D-dimer is a specific degradation product of cross-linked fibrin. Plasma concentration is raised in thromboembolism, reflecting fibrin formation and dissolution, so measurement of plasma D-dimer has a potential use in the diagnosis of DVT. Monoclonal antibodies against neoantigens exposed in the split products of fibrin allow direct plasma assay.Methods for testing D-dimer include ELISA, latex agglutination and immunofiltration. ELISA has consistently shown high sensitivity: 98% (Dale et al, 1994), 96 . 8% (Bournameux et al, 1994) and negative predictive values: 95% (Dale et al, 1994) for the diagnosis of DVT. Low specificity for predicting DVT: 38% (Dale et al, 1994), 35 . 2% (Bournameux et al, 1994 and positive predictive value: 54% (Dale et al, 1994), limits D-dimer to a test of exclusion only. Although accurate, ELISA is time consuming and is not a realistic choice for a screening test in an emergency or outpatient setting. Latex agglutination is faster but only has a sensitivity of 73% for the diagnosis of DVT (Dale et al, 1994), making it unsuitable as an exclusion test.NYCOCARD D-dimer (Nycomed Pharma AS, Norway) is an immunofiltration system. Preliminary results suggest that the sensitivity of the NycoCard system for D-dimer approaches that of ELISA. (Dale et al, 1994). The assay time is less than 2 min, so the test is suitable for use in acute care. PATIENTS AND METHODSThe study was designed as a double-blind prospective investigation of the use of the NycoCard D-dimer test in patients with suspected DVT. Entry criteria for the study population were medical inpatients or outpatients referred to the St Helier ...