Testing for d-dimer to exclude venous thromboembolic disease has been available since the 1980s, but the clinical role of this test remains unclear because of the variety of available assays and concerns about their sensitivities and specificities. Stein and associates reviewed the available evidence on each of the various d-dimer assays.

Using prospective studies of controlled populations where sensitivity and specificity data were present or could be calculated, studies were grouped by the type of d-dimer assay used, the cutoff points for positive tests, and whether d-dimer was used to exclude pulmonary embolism (PE) or deep venous thrombosis (DVT).

DVT analyses found that, in terms of sensitivity, the enzyme-linked immunosorbent assay (ELISA) was clinically and statistically superior to the the quantitative latex agglutination and semi-quantitative latex agglutination assays. For specificity, the quantitative latex agglutination and whole-blood agglutination tests were clinically and statistically superior to the ELISA assay. ELISA testing had a high negative likelihood ratio, which give a high certainty for a negative diagnosis. The estimated positive likelihood ratios for most of the tests are poor; a positive test does not increase the certainty of a positive diagnosis. PE analyses found similar clinical and statistical characteristics as the DVT analyses, with the ELISA assays having strong negative likelihood ratios. The most commonly examined d-dimer cutoff value was 500 ng per mL, although some data were available for cutoff values of 250 ng per mL and 1,000 ng per mL.

The authors conclude that the ELISA and qualitative rapid ELISA tests have the least variability and best sensitivity and ability to rule out DVT and PE. The specificity values of the various tests were not adequate to have clinical value in altering the probability of disease, making d-dimer testing a unidirectional test: a negative result is useful to exclude DVT or PE, but a positive result does not necessarily indicate the presence of disease. The negative ELISA assay provides a negative likelihood ratio similar to those of a normal perfusion lung scan or a negative Doppler ultrasonography finding, making the ELISA assay usable as a stand-alone test. The other assays do not have the same accuracy. Certainly, final diagnostic accuracy is enhanced by using the d-dimer assay in a diagnostic pathway that also identifies pre-testing clinical probability for DVT or PE.

In a commentary in the same issue, Sox points out the importance of determining the pretest probability of DVT to optimize the d-dimer assay result. In patients with a high pretest probability of disease, a negative d-dimer test may not be enough to support forgoing anticoagulation therapy.