The measurement of D-dimers is a recent addition to the diagnostic strategy of pulmonary embolism and has been shown to be a valuable tool with excellent sensitivity. However, there have been rare reports of patients with pulmonary embolism but negative D-dimer tests. The object of this study was to study patients with pulmonary embolism but negative D-dimers and to compare them with a population of patients with pulmonary embolism and raised D-dimers. One hundred and fifty consecutive patients admitted for pulmonary embolism were included in this study. All underwent measurement of D-dimers (normal <500 ng/ml) by an ELISA technique. The data of clinical examination and complementary investigations were analysed with respect to the D-dimers result. The sensitivity of raised D-dimers for pulmonary embolism was 96% (6 patients had results <500 ng/ml). The finding of chest pain was statistically greater in the group with negative D-dimers (p=0.01). In these cases, the emboli were all distal (p=0.0003), the average Miller index was significantly lower than in patients with high D-dimers (p=0.04) and the diagnostic value of ultrasound investigations (echocardiography, ultrasonography of lower limb veins) was less (p<0.0001). The authors conclude that measurement of D-dimers by the ELISA method may be non-diagnostic in distal pulmonary embolism and one explanation could be the less extensive thromboembolic process. In cases with negative D-dimers, a strong clinical suspicion of pulmonary embolism should lead to the request for further investigations.