D-dimer testing in the diagnosis of acute venous thromboembolism

Abstract

Patients with acute VTE require clinical assessment and objective testing to be accurately diagnosed. Almost all patients with acute VTE have an elevated D-dimer level. An elevated D-dimer is associated with many illnesses, and therefore, is not specific for VTE. D-dimer tests can have a high sensitivity, however, which is useful because a normal test excludes the diagnosis of VTE. D-dimer testing is most appropriate in the assessment of outpatients because the prevalence of disease and the likelihood of comorbid conditions are lower than in inpatient populations, making a test of exclusion particularly valuable. Accuracy studies using conventional ELISA assays have confirmed that a test with a high sensitivity can be used to exclude a diagnosis of VTE, but conventional ELISA testing is not practical. Studies of more practical D-dimer testing indicate that, for patients with suspected DVT or PE, the need for serial testing or further investigation can be reduced if normal results are obtained using assays with a high sensitivity. There are, however, many sources of variation in the test characteristics of D-dimer assays. Therefore there is no reassurance that results from one manufacturer's test are applicable to other tests and different investigators may obtain varied results when using the same manufacturer's product. In addition, the results of D-dimer accuracy studies lack generalizability. This lack of generalizability has led to the recommendation that clinicians await the results of management studies before adopting the routine use of D-dimer assays in the diagnosis of VTE. Further, it may be reasonable to perform an accuracy study when planning to adopt a specific D-dimer assay from a published management trial, to be confident of its characteristics can be reproduced. In the management of patients with suspected DVT, rapid ELISA tests show promise as a practical D-dimer test, in that they have a sensitivity similar to that of the conventional ELISA assay. Two management studies have recently confirmed that a normal D-dimer result (using the SimpliRED whole-blood assay or the Instant IA rapid ELISA) in combination with a noninvasive test or a clinical model can reliably exclude DVT in outpatients. Use of a clinical model can reduce the need for VU, and the combination of a clinical model and D-dimer testing could further reduce the number of VU procedures required. As noted by Wells et al, who recently published a clinical model, however, a normal D-dimer result was most accurate in the patients with a low pretest likelihood (NPV = 99.5%) and least accurate in patients with a high pretest likelihood (NPV = 85.7%) Patients with a low pretest likelihood and a normal D-dimer are the largest proportion of outpatients referred for testing, and considerable resources may be saved if additional management studies confirm the usefulness of D-dimer testing in such patients. In patients with suspected PE, there is a lack of published management trials despite a number of accuracy studies indicating that D-dimer testing may be useful as a method of PE diagnosis exclusion. Recent results, however, from an accuracy study of patients with suspected PE who had D-dimer testing complement the findings of Wells et al in patients with suspected DVT. Using a standardized clinical model of PE in combination with a SimpliRED D-dimer assay, Ginsberg and colleagues found that the combination of a low pretest likelihood and a normal D-dimer had a negative predictive value of 99%, whereas the negative predictive value was only 78% in patients with a high pretest likelihood and a normal D-dimer. Similar to the findings in DVT, these results indicate that D-dimer testing is most useful in patients with a low pretest likelihood for PE and raise the possibility that such patients may not require lung scans. This finding is currently being evaluated in a prospective management trial.