Purpose: Large studies have shown that most cases referred for duplex sonography for suspected deep vein thrombosis (DVT) have normal scan results. For medical and economic reasons, a preselection procedure, which allows the detection of true-negative cases before duplex scanning, is required; this procedure should be characterized by a high sensitivity and a high negative predictive value.
Methods: In 343 patients (398 lower extremities) with suspected DVT, the DVT probability was clinically assessed, and a whole blood D-dimer agglutination test and a duplex scan were performed. The diagnostic sensitivities of the D-dimer test alone, a high clinical DVT probability alone, and the combination of both were evaluated.
Results: The sensitivity values for the D-dimer test to diagnose proximal and distal DVTs were 88.7% and 80.9%, the negative predictive values (NPV) were 96.3% and 97.9%, and the specificity and the positive predictive value (PPV) were 54.8% and 49.6% and 26.6% and 8.2%, respectively. The sensitivities of the clinical DVT probability assessment for the diagnosis of proximal and distal DVTs were 83.9% and 66.7%, respectively; the corresponding NPVs were 94.9% and 96.5%, respectively. The specificity was 56.1% and 50.8%, and the PPVs were 26.1% and 7.0%, respectively. The combined use of the results of the clinical probability assessment and the D-dimer test resulted in sensitivities for proximal and distal DVTs of 98.4% and 90.5%, NPVs of 99.3% and 98.6%, a specificity of 43.4% and 38.4%, and PPVs of 24. 3% and 7.6%, respectively.
Conclusion: The combined use of a clinical DVT probability assessment scheme and the D-dimer test largely avoids false negative results, has a high sensitivity and NPV, helps to reduce the costs of DVT diagnosis, and may, in the future, be useful as a preselection procedure before duplex sonography.