Why the study?
Does D-dimer testing accurately diagnose or exclude venous thromboembolism in patients with suspected disease?
Does D-dimer testing accurately diagnose or exclude venous thromboembolism in patients with suspected disease?
Highly sensitive D-dimer assays are clinically useful for excluding venous thromboembolism, but their low specificity requires them to be used selectively in appropriate patient populations with low pre-test probability.
Answers are grounded in this paper and cite their sources.
Supports D-dimer rule-out in low-probability VTE; leaves open assay performance across broader populations.
The diagnosis of thrombotic disorders has received a great deal of attention in the medical literature over the past decade. The diagnostic modalities of both distal venous thrombosis (DVT) and pulmonary embolus (PE), collectively referred to as venous thromboembolism (VTE), are salient to the practice of emergency medicine because these diseases are frequently in the differential diagnosis of patients presenting to the emergency department. This concern is compounded by the fact that up to 60 percent of patients with DVT will develop a pulmonary embolism if not treated.1–4 The morbidity and mortality of this potential complication are of major concern because PE accounts for about 250,000 hospitalizations and as many as 50,000 deaths per year in the United States, and of these, an estimated 30 percent are first seen in the ED.5–7 Despite these alarming numbers, the diagnosis of these disorders remains challenging. For DVT, clinical findings are clearly unreliable; only 30 percent of all patients with suggestive clinical findings referred to the ED to rule out this diagnosis actually have a DVT.8–10 The diagnosis of PE is equally difficult, and the consequences of missed diagnoses can be devastating. The diagnosis of PE is missed more than 400,000 times per year in the United States, and approximately 100,000 patients who would have survived with the proper diagnosis die annually.1 Because of the high incidence of VTE, the large numbers of misdiagnoses, and its significant morbidity and mortality, numerous strategies have been developed to help the clinician make timely and cost-effective decisions. These include invasive and non-invasive testing modalities, each with their own sensitivities, specificities, risks, and benefits. D-Dimer A novel approach to the diagnosis of VTE that has received a great deal of scrutiny recently is the use of a D-dimer assay. D-dimer is a degradation product of fibrin that is formed when fibrin is degraded by plasmin. Circulating levels can be detected as early as one hour after thrombus formation.11–13 Although D-dimers have a circulating half-life of four to six hours, the continued fibrinolysis of a clot (either DVT or PE) increases the plasma D-dimer concentration for at least one week.14,15 Other studies suggest, however, that false-negative test results may result in patients with symptoms of thrombotic disease lasting longer than seven days because circulating D-dimer concentrations may normalize after this time.16–18 The specificity of the assay is limited by the fact that D-dimer levels can be elevated in many other conditions such as sepsis, recent myocardial infarction or stroke (within 10 days), recent surgery or trauma, disseminated intravascular coagulation, pregnancy or recent delivery, active collagen vascular disease, metastatic cancer, and liver disease. Because of the high false-positive result rate of these tests, they are used primarily to exclude the diagnosis of VTE.Table: D-Dimer AssaysTesting Methods Widespread use of D-dimer testing is relatively new, and to date, no single test has shown the requisite combination of sensitivity, specificity, timeliness, and low cost that would make it practical for use in the ED. To understand the benefits and limitations of the various tests, it is important to examine the most commonly used and best-studied tests individually. There are three main methods used for D-dimer detection (ELISA, whole blood agglutination, and latex agglutination). The first, which is used as the reference standard to which the other assays are compared, is the enzyme-linked immunosorbent assay (ELISA). This quantitative test is expensive and time-consuming, making it impractical for the ED diagnosis of thrombotic disease. Rapid ELISA tests are available, both quantitative and qualitative, which provide a result within 35 minutes, but the sensitivity of these tests is inconsistent. Additionally, there are a number of agglutination tests (latex, immunofiltration, and whole blood) that are rapid and inexpensive but do not have a consistent sensitivity comparable with the ELISA. The whole blood agglutination technique (SimpliRED, Agen Diagnostics Limited) uses a biospecific antibody with epitopes that are portions of D-dimer and red blood cells, respectively. In the presence of elevated levels of D-dimer, the antibody causes agglutination of the patient's red blood cells. The benefits of this type of assay are that it is fast, inexpensive, and can be done at the bedside. Clinical Algorithm Numerous studies have assessed the quality of these tests, and some have even attempted to implement them in a clinical algorithm for the detection of thrombotic disease.16,19,20–22 Freyburger et al studied 100 consecutive patients with clinically suspected DVT (n=83) or PE (n=17) with eight different D-dimer methods in a blinded fashion. Each patient underwent bilateral ascending contrast venography and intra-arterial pulmonary angiography with selective catheterization of each pulmonary artery. The results of the gold standard tests were then compared with the various D-dimer method used. The sensitivities of the ELISA tests (Asserachron D-Di Stago, D-dimer micro Behring, and Diemertest Gold EIA Agen) were 100% while the specificities ranged from 34% to 52%. Only the rapid quantitative VIDAS (bioMérieux, Marcy l'Etoile, France) assay was found to have a sensitivity of 100% and a specificity of 38%. This was the only method comparable with the standard ELISA.19 Of note, no distinction in the overall sensitivities and specificities was made between PE and DVT. These data simply denote the presence or absence of clot. While the results of this study are promising, especially for the rapid ELISA, its results are somewhat limited by the small population size and even smaller number who were proven to have the diseases in question by gold standard testing. For instance, of the 83 clinically suspected cases of DVT, only 43 patients were proven to have the disease (true positive), and of the 17 cases suspected of PE, only seven patients had angiographically proven disease. In a study of 99 consecutive outpatients with suspected DVT, van der Graaf et al compared 13 D-dimer methods with contrast venography including six novel quantitative assays and four agglutination assays (both latex and whole blood) as well as three standard ELISA methods as reference standards. The cohort of patients was comprised of consecutive outpatients presenting to the emergency department with clinical suspicion for DVT. The results of the study indicate that of the six novel quantitative assays evaluated, the VIDAS and Tinaquant showed the highest sensitivities (100% for each). The qualitative Minutex and SimpliRED assays demonstrated a relatively high specificity but inadequate sensitivity for the diagnosis of DVT (90%/80% and 94/80%, respectively).22 D-dimer testing is useful in the clinical decision-making process, particularly in the appropriate patient population Meyer et al investigated the sensitivity of semiquantitative enzyme-linked immunosorbent assay (Instant IA D-Dimer), a quantitative assay (STA-Liatest DDi) as compared with standard ELISA in the diagnosis of clinically suspected pulmonary embolism in 142 patients. The sensitivities of ELISA and STA-Liatest DDi were 92% [95% confidence interval (CI) 82–97%] and 93% (95% CI 84–98%), respectively, while Instant IA D-dimer sensitivity varied from 83% to 93%. This study is somewhat limited by the wide confidence intervals, and the lack of use of a gold standard for the diagnosis of PE (angiography) in all patients in the study. The authors conclude, however, that STA-Liatest DDi is comparable with standard ELISA for use in exclusion of pulmonary embolism.21 Oger et al also investigated the sensitivity of the STA-Liatest DDi in the diagnosis of pulmonary embolus. In their study of 386 consecutive patients suspected of pulmonary embolus, they found the Liatest to be 100% sensitive. Of note however, is that, again, no consistent gold standard test was used to determine the presence of PE. Diagnosis of PE was based on “clinical evaluation, lower limb examination, radionuclide lung imaging, and, when required, pulmonary angiography.”16 No Consensus Although numerous studies have assessed the use of D-dimer testing in the presence of VTE, there has been no consensus on its inclusion in the diagnostic evaluation of the ED patient. Part of the reason for this is the fact that not all D-dimer testing is equal. The range of sensitivities varies with the type of assay used, which is an important distinction if D-dimer testing is the sole test used to rule out thrombotic disease. Although standard ELISA may show the requisite sensitivity for its use as a “rule-out” test, the cost and lack of timeliness may limit its appropriateness in the ED. To counter this problem, new rapid ELISA assays were developed, but as can be seen by the variation in sensitivities, not all assays are equal. The clinician must know which test is used in his laboratory and understand its strengths and weaknesses. Other novel qualitative approaches have been developed, such as the whole blood and latex agglutination assays, in an attempt to create a test that is applicable for the ED setting. These have been shown to be fast and easy to use, but their sensitivities are not high enough to be used as a sole “rule-out” test. The varied findings of these studies may result from the heterogeneity of the populations. Prevalence of disease, duration of illness, and comorbid diseases all serve to confound the data. For instance, negative predictive value (NPV) is frequently used as high-yield criteria for examination of exclusionary testing. Because NPV is dependent on the prevalence of disease in the population studied, it is difficult to compare different tests used in different populations with different disease prevalence. For example, with a prevalence of DVT of 50 percent, a sensitivity of 98% and a specificity of 40%, the NPV would be 95%. With the same sensitivity and specificity but a prevalence of only 30 percent, the NPV would be almost 99%.22 Therefore, one may increase the effectiveness of the screening test by using it selectively in patients who have a low pre-test probability. Of note, the specificity of all tests studied is universally low. Because there are many conditions that can yield a positive result and the population in which this test will be used will likely harbor many of the conditions in question, the clinician will more than likely be forced to continue down the diagnostic pathway with another test. Additionally, DVT and PE may not be equal diseases when it comes to D-dimer testing. There are many conditions that affect the presence of D-dimer which may explain the invariably low specificities noted in all assays studied. Chief among these is clot burden. It is quite possible that the low clot burden required for pulmonary embolus, and therefore, the low levels of circulating D-dimer, make detection of PE with D-dimer testing more difficult. It appears that the sensitivity of appropriate D-dimer testing is not adequate for its sole use in the exclusion of PE. Van Beek calculated that for every two percent decrease in sensitivity one per 1000 evaluated with clinically suspected people would die.23 This does not mean, however, that it does not have its place in the diagnostic evaluation of PE. It would seem appropriate that the use of D-dimer would be one more piece of information that the clinician could use to aid in the decision process. D-dimer use in patients at higher risk for VTE or with confounding conditions is only an additional piece of information in the diagnostic work-up This approach has been investigated by Ginsberg et al in his study of 1177 consecutive patients suspected of having a pulmonary embolus. All patients in the study underwent D-dimer testing, ventilation-perfusion scanning, and bilateral compression ultrasonography. Patients in whom PE was not diagnosed were followed for three months. Of the patients studied, 703 were categorized as low pretest probability. In this subset, a negative D-dimer was found to have a likelihood ratio of 0.27, for a posterior probability of PE of 1.0%.24 From these data, the authors conclude that a normal D-dimer is useful to exclude pulmonary embolus in patients with low pretest probability. The numbers in this study make it attractive, but again, the accepted gold standard, angiography, was not applied universally in this study. Heit et al also prospectively studied the use of D-dimer testing in the face of suspected PE. One hundred and five patients with a clinical suspicion for PE were tested with pulmonary angiography and D-dimer assays, and 33 (31%) had a positive pulmonary angiogram. The D-dimer sensitivity/negative predictive value for the ELISA, latex agglutination (American Bioproducts Co./Diagnostica Stago and Biopool International), and membrane ELISA were 100%/100%, 94%/94%, 100%/100%, and 97%/96%, respectively. Sensitivities and specificities were calculated at a discriminate level of the assays of 250 microg/L or less. The authors concluded that D-dimer is a sensitive but not specific test for the diagnosis of PE at the appropriate discriminate level.25 Finally, Brown et al performed a meta-analysis of 52 clinical trials using ELISA D-dimer testing in the diagnosis of pulmonary embolus, nine of which met inclusion criteria in their study (prospective, ELISA, and outpatient population). In their analysis, ELISA had a sensitivity of 0.94 (95% CI 0.88-.097) and a specificity of 0.45 (95% CI 0.36 to 0.55). The authors concluded that the test might help clinicians safely rule out PE, especially in the face of low and low-to-moderate pretest probabilities.26 While the results of this meta-analysis would seem to make ELISA attractive for use in ruling out the presence of disease, ELISA testing is not rapidly available at most institutions, if at all. Therefore, the utility of this study to the practicing emergency physician could be limited. Venous thromboembolism is a clinically important disease that can be diagnostically challenging. Numerous invasive and non-invasive diagnostic methods have been devised, one of which is D-dimer testing. Based on available data, it would seem that use of this test as an exclusive screening test is highly dependent on the type of test used, the pre-test probability of disease, and the type of VTE suspected (DVT vs. PE). Based on the available data, D-dimer testing should be considered a useful piece of information in the clinical decision-making process, particularly in the appropriate patient population. When used in patients with a low pre-test probability of VTE and a lack of comorbid diseases or other confounding conditions, the ELISA D-dimer test appears to be a safe and effective tool in ruling out VTE. However, its use in patients at higher risk for VTE or with confounding conditions can only be advocated as an additional piece of information in the diagnostic work-up.
No takes yet. Share an insight, caveat, or question.
Mosely et al. (2003) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: