ABSTRACT
BACKGROUND: The diagnosis of a pulmonary embolism is challenging because of the potentially severe consequences of missing the condition and imaging test results that are often equivocal. Clinical decisions may be aided by a test for D-dimer, a fibrin degradation product usually increased in the blood of patients with thromboembolic disease. Laboratory advances have led to more rapid and practical detection techniques, such as the enzyme-linked immunosorbent assay (ELISA) method examined by this meta-analysis.
POPULATION STUDIED: The 11 prospective studies included in this meta-analysis involved 2126 patients (aged 54 to 81 years), more than 98% of whom were outpatients presenting with symptoms and signs suspicious for a pulmonary embolism. There were more women than men in the study population.
STUDY DESIGN AND VALIDITY: MEDLINE and EMBASE searches identified relevant titles and abstracts of published articles from 1980 through 2000, and the researchers also searched for unpublished work, screened the reference lists of articles, and contacted authors to identify other studies. All of the included studies had to be original prospective investigations of the ELISA D-dimer test enrolling at least 80% outpatients. Independent reviewers screened articles for quality of reference standards and generalizability. Reference standards for the diagnosis of a pulmonary embolism included a high-probability ventilation/perfusion scan, positive computerized tomographic scan, or positive lower extremity imaging. Acceptable standards for a negative diagnosis were normal or very low probability ventilation/perfusion scan or the absence of a thromboembolic event for 3 months or more.
OUTCOMES MEASURED: The primary results were the pooled sensitivity and specificity of the ELISA D-dimer test in detecting a pulmonary embolism. The authors reported pooled estimates of sensitivity and specificity for various combinations of studies, based on the quality of the studies, method of ELISA testing, age of patients, comorbid conditions, and symptom duration.
RESULTS: Pooling all 11 studies led to an overall sensitivity of 0.95 (95% confidence interval [CI], 0.90–0.98) and a specificity of 0.45 (95% CI, 0.90–0.98). The 5 studies that used the most rigorous reference standard protocols yielded a sensitivity of 0.90 and a specificity of 0.40. A higher pooled accuracy was found in the group of 8 studies with the most typical spectrum of outpatients (sensitivity and specificity of 0.97 and 0.48, respectively). In general, the high sensitivity and low specificity of the ELISA D-dimer test will lead to the detection of a pulmonary embolism most of the time when the condition is present, but false-positive results are common. Several clinical scenarios may decrease the test’s accuracy. In the one small study of patients aged 70 years or older, the sensitivity was perfect (1.00) but the specificity was low (0.14). In another study examining patients whose symptoms lasted 4 days or more, the sensitivity fell to 0.73 and specificity was 0.33. Both of these studies suggest more consideration be given to other conditions, such as infection, cancer, or inflammatory arthritides, which could elevate D-dimer levels in these patient populations.
The ELISA D-dimer test gives relatively reliable information to rule out an acute pulmonary embolism in outpatients. A negative ELISA D-dimer result is particularly useful in ruling out a pulmonary embolism when the pretest probability of a pulmonary embolism is low, and a clinical assessment tool can help determine who fits such a low-probability profile.1 Patients with a higher pretest probability should probably still have an imaging test, considering the morbidity and mortality of a missed pulmonary embolism. Confirmatory imaging is also indicated in most instances when the D-dimer test is positive, due to the poor specificity of the test.