Predictive value of a rapid semiquantitative D-dimer assay in critically ill patients with suspected venous thromboembolic disease

Objective: To evaluate the performance of a new, rapid semiquantitative assay for the detection of circulating D-dimer in whole blood from critically ill patients with suspected venous thromboembolic disease. Design: Prospective, blinded, single-center study. Setting: Medical intensive care unit (ICU) of Barnes-Jewish Hospital, St. Louis, MO, a university-affiliated urban teaching hospital. Patients: Two hundred thirty-nine adult patients with clinical suspicion of venous thromboembolic disease admitted to a medical ICU. Interventions: Collection of blood samples within 24 hrs of clinical suspicion of venous thromboembolic disease. Measurements and Main Results: The main outcome measures evaluated included the occurrence of venous thromboembolic disease (i.e., lower extremity venous thrombosis, pulmonary embolism, catheter-associated venous thrombosis) and hospital mortality. Fifty-seven patients (23.8%) were classified as having venous thromboembolic disease during their ICU stays (pulmonary embolism, 21 patients; lower extremity thrombosis, 44 patients; line-associated venous thrombosis, 3 patients). The semiquantitative whole-blood assay for circulating D-dimer had a 96.4% sensitivity and a negative predictive value of 92.1% for identifying patients with venous thromboembolic disease. The specificity of this assay was 19.7%, and its positive predictive value was 26.9%. There was a strong correlation between the semiquantitative assay for circulating D-dimer and the quantitative measurement of circulating D-dimer using an enzyme immunoassay (Spearman's correlation coefficient, 0.8643; p < .001). We also identified a strong correlation between both the quantitative and semiquantitative measurements of circulating D-dimer with the sepsis classification proposed by the American College of Chest Physicians/Society of Critical Care Medicine (i.e., systemic inflammatory response syndrome, sepsis, severe sepsis, septic shock) for patients without venous thromboembolic disease (n = 182; quantitative measure: Spearman's correlation coefficient, 0.207; p = .002; semiquantitative measure: Spearman's correlation coefficient, 0.3519; p < .001). Conclusions: These preliminary findings suggest that a rapid whole-blood assay for the semiquantitative detection of circulating D-dimer may be a useful diagnostic tool for the exclusion of venous thromboembolic disease among critically ill patients.