The true relationship between plasma D-dimers and acute ischaemic stroke (AIS) is uncertain as previous studies investigating this have not screened for subclinical deep vein thrombosis. We addressed this as part of a study in which we screened AIS patients for venous thromboembolism (VTE). We also assessed the performance characteristics of two D-dimer assays as exclusionary tests for VTE in these patients. One hundred and two unselected AIS patients were screened for VTE using magnetic resonance direct thrombus imaging. D-dimers were analysed on days 2, 9, 14 and 21 using the VIDAS immunofluorescent assay (cut-off >or= 500 ng/ml) and the IL test D-dimer immunoturbidimetric assay (cut-off >or= 255 ng/ml). The relationship between D-dimers and AIS was examined in 52 patients neither developing VTE nor intercurrent illness. D-dimers were elevated throughout the study. Median values at the four time points were 652, 692, 737 and 686 ng/ml (VIDAS assay) and 260.5, 268.5, 273 and 283 ng/ml (IL assay). D-dimers were higher in patients aged older than 70 years, with severe stroke or with total anterior circulation infarcts: only age older than 70 years was significantly associated with D-dimer values greater than the median on univariate and multivariable analysis. Both assays were 100% sensitive for VTE. Specificities were 30% (VIDAS assay) and 34% (IL assay). Specificity was adversely affected by age older than 70 and severe versus non-severe stroke. D-dimers are elevated in the first 3 weeks post-AIS after eliminating the confounding effect of subclinical deep vein thrombosis. The VIDAS and IL assays remained sensitive tests for VTE but the specificity was low, limiting their exclusionary efficiency in these patients.