대한내과학회지

BACKGROUND/AIMS: The D-dimer value is a simple blood test used to evaluate venous thromboembolism (VTE). However, due to its low specificity, another test is needed for a definite diagnosis, such as a radiographic test. We evaluate the factors associated with a false positive D-dimer test and propose a new cut-off value for detecting VTE more effectively in Koreans. METHODS: This was a retrospective, observational study. From January 2009 to December 2009, 2,047 patients (988 men, 63 +/- 15 years) had the D-dimer value checked to evaluate VTE. The main outcome of interest was a positive D-dimer test. Odds ratio and 95% confidence intervals were determined using logistic regression analysis. The new D-dimer cut-off was evaluated using receiver operating characteristics (ROC) curves. RESULTS: The result was positive in 1,093 patients (53%), for a false positive percentage for VTE of 95% and a false negative percentage for VTE of 1%. Significant false positive predictors for a positive D-dimer were increasing age, trauma, postoperative, acute infection, tuberculosis, stroke, malignancy, chronic renal failure, acute coronary syndrome, heart failure, and lung disease. The discriminative value of the D-dimer test was assessed using ROC curve analysis. A D-dimer value of 0.68 mg/L on admission was the best cut-off value for predicting the development of VTE with a sensitivity of 95% and specificity of 57%. CONCLUSIONS: Many factors affect the D-dimer value and we must consider these factors before using the D-dimer value to evaluate VTE. A D-dimer value of 0.68 mg/L appears to be a good cut-off value for evaluating VTE more effectively in Koreans.