We studied the usefulness of the determination of plasma D-dimer levels (using an ELISA) in combination with non-invasive testing with impedance plethysmography (IPG) or real-time ultrasonography (US) for the diagnosis of deep-vein thrombosis (DVT), in outpatients with clinically suspected DVT. This combined approach was compared to serial non-invasive testing alone in these patients. The sensitivity of a positive D-dimer test (greater than 300 micrograms/l) for the presence of DVT was 100% (70/70 patients; 95% C.I.: 95-100%), whereas the specificity was 29% (69/239 patients; 95% C.I.: 23-34%). The proportion of patients in which a definitive decision about the presence or absence of DVT could be made on the day of referral, was calculated for both approaches. When applying the combined approach, in 42% of all referred patients the diagnosis of DVT could either be established or refuted on entry, as opposed to only 19% of patients using serial non-invasive testing alone. Also, the costs per DVT diagnosed were calculated for the two diagnostic approaches. For the diagnosis of DVT the costs using serial IPG were comparable to the costs using the combination of IPG and the D-dimer test. The same conclusion holds for the comparison of serial US with the combination of US and D-dimer testing. We conclude that for the diagnosis of DVT in symptomatic outpatients the combination of non-invasive testing with the D-dimer test might be preferred over serial non-invasive testing alone, although the safety of such an approach remains to be established in future management studies.