A normal D-dimer (DD) concentration for the exclusion of deep venous thrombosis (DVT) has a low specificity in older patients and compression ultrasonography is often required. Three D-dimer assays, STA Liatest, Tina-quant, and Innovance, are evaluated in symptomatic outpatients suspected for DVT with emphasis on its performance in older patients by using different cut-off levels. This study includes 466 outpatients suspected for having DVT. The diagnostic accuracy, measured as sensitivity and area under the curve of the receiver operation characteristic curve is good for all DD assays. The specificity of the DD assays combined with a low pretest probability varies from 42.6 to 51.5%. The specificity of the three DD assays in patients > or = 60 years varies, however, between 24.6 and 40.9%. Several cut-off values in different age-subgroups are studied. For patients < 60 years, the most accurate cut-off value is 500 microg/L for all DD assays. For patients > or = 60 years, a threshold of 750 microg/L has the best results with NPV of 100% for all assays and specificity of 48.5% (STA Liatest), 60.6% (Tina-quant), and 49.2% (Innovance), respectively. For the three assays, the number needed to test (NNT) decreases in both subgroups of patients compared to the standard algorithm. A cut-off level of 750 microg/L for patients > or = 60 years improves the clinical performance of DD assays in combination with the PTP score without the loss of NPV. The NNT improves substantially with an age-adapted algorithm.