OBJECTIVE Several clinical studies have proposed using d-dimer as an initial screening test to exclude thrombosis in cases of suspected (DVT). In published series, these assays have variable sensitivity, raising concerns that they may not be sufficiently robust for clinical practice. The aim of the study was to examine the sensitivity of two commercially available d-dimer assays and to assess their value and safety as initial screening tests in suspected DVT. METHODS In this prospective study, blood samples were collected for d-dimer measurement (SimpliRED assay and IL test d-dimer) in all patients presenting to the emergency department over a twelve month period. All patients underwent compression ultrasound scanning as the primary diagnostic procedure. RESULTS 235 patients were included in the study. 51(22.8%) had a DVT confirmed on ultrasound. The SimpliRED assay was positive in only 33 cases, with seventeen cases of confirmed DVT giving a negative result (six cases with proximal vein thrombosis). Assay sensitivity was 66%, with a negative predictive value of 88.9%. The IL test gave three false negatives (all below knee thromboses) giving assay sensitivity and negative predictive value of 94.1% and 96.8% respectively. CONCLUSIONS The precise role of d-dimer testing in the diagnosis of venous thrombosis has yet to be established. From our results and a review of published series, we conclude that the SimpliRED assay is too insensitive to use as a reliable exclusion test in cases of suspected DVT, however, the more sensitive automated IL test d-dimer may have a role in the initial assessment. We propose that the IL d-dimer test is used in conjunction with a pre-test probability score to identify patients at low risk of DVT and recommend that this approach is tested in a clinical study before introduction into practice.