Diagnosis of deep vein thrombosis: how many tests do we need?Acta Chir Belg. 2005 Feb; 105(1):16-25.AC
The requirement for a safe diagnostic strategy should be based on an overall post-test incidence of venous thromboembolism (VTE) of less than 1% during 3 month follow-up. Compression ultrasonography (CUS) has a negative predictive value (NPV) of 97 to 98% indicating a post-CUS incidence of deep vein thrombosis (DVT) of 2 to 3%. A post-CUS DVT incidence of 3% implicates that 90 to 120 DVTs per 1 million inhabitants will be overlooked each year indicating the need to improve the diagnostic work-up of DVT as much as possible. The qualitative D-dimer test (SimpliRed) has a sensitivity of 82 to 89% and a negative predictive value of 94 to 95% indicating a 5 to 6% post-test incidence of DVT, which is not sensitive enough for venous thrombosis exclusion. The post-test DVT incidence could be reduced from 3.2% to 0.6% in one study and from 11% to 2% in another study by the combination of a normal CUS and low clinical score and from 4.5% to 1.6% by the combination of low clinical score and a negative SimpliRed test in one study. The combination of a negative CUS and a negative SimpliRed test reduced the post-test incidence of DVT from 2.6% to < 1% or even < 1% in two management studies without the need of a repeated CUS on the basis of which anticoagulant therapy can safely be withheld. The rapid quantitative turbidimetric D-dimer assay (Tinaquant) has a sensitivity and a negative predictive value (NPV) of 97.7% with a 2.3% post-test incidence of DVT. The combination of a normal Tinaquant D-Dimer test result plus a low to moderate clinical score reduces the post-test incidence of DVT from 2.3 to 0.6% without the need of CUS testing in 29% of patients with suspected DVT. The rapid ELISA VIDAS D-dimer assay has a sensitivity and NPV of 98.6 and 99.5% in two management studies for the exclusion of DVT irrespective of clinical score. The combination of a normal ELISA VIDAS D-Dimer test with clinical score assessment will reduce the post-test DVT incidence of less than 0.5% and the need for CUS testing by 40 to 50%. It is concluded that the sequential use of a rapid quantitative D-dimer test, clinical score and CUS appears to be safe and the most cost-effective diagnostic work-up of DVT.