A comparative analysis of D-dimer assays in patients with clinically suspected pulmonary embolism

To avoid angiography in patients with clinically suspected pulmonary embolism and non-diagnostic lung scan results, the use of D-dimer has been advocated. We assessed plasma samples of 151 consecutive patients with clinically suspected pulmonary embolism. Lung scan results were: normal (43), high probability (48) and non-diagnostic (60; angiography performed in 43; 12 pulmonary emboli). Reproducibility, cut-off values, specificity, and percentage of patients in whom angiography could be avoided (with sensitivity 100%) were determined for two latex and four ELISA assays. The latex methods (cut-off 500 micrograms/l) agreed with corresponding ELISA tests in 83% (15% normal latex, abnormal ELISA) and 81% (7% normal latex, abnormal ELISA). ELISA methods showed considerable within- (2-17%) and between-assay variation (12-26%). Cut-off values were 25 micrograms/l (Behring), 50 micrograms/l (Agen), 300 micrograms/l (Stago) and 550 micrograms/l (Organon). Specificity was 14-38%; in 4-15% of patients angiography could be avoided. We conclude that latex D-dimer assays appear not useful, whereas ELISA methods may be of limited value in the exclusion of pulmonary embolism.