The D-Dimer Test for Deep Venous Thrombosis: Gold Standards and Bias in Negative Predictive Value (Evidence-Based Laboratory Medicine and Test Utilization‪)‬

The measurement of d-dimer, a degradation product of circulating cross-linked fibrin, has attracted much attention as a rule-out test for venous thromboembolism. In the past 20 years, more than 40 studies have been published comparing d-dimer concentrations with a gold standard for lower extremity deep venous thrombosis (DVT) (129). Until 1995, d-dimer studies usually used lower extremity venography as the gold standard for DVT diagnosis (1). More recently, paralleling clinical practice, duplex ultrasonography has assumed a prominent role as the gold standard. In 28 studies published between March 1995 and October 2001, duplex ultrasound alone was used as the gold standard in 9 studies (2-5, 7, 8, 11, 14, 17), ultrasound and/or venography in 9 studies (6, 10, 12, 18, 25, 26, 28, 29), and venography alone in 10 studies (13,15, 16, 19-22, 24, 27, 30). Although clinicians now use duplex ultrasound as the test of choice for the initial evaluation of patients with suspected lower extremity DVT (31), ultrasound has disadvantages as a gold standard for lower extremity DVT research because of its limited ability to image calf DVT (32). In clinical practice, this limitation has been circumvented through the use of serial testing strategies designed to detect propagation of calf DVT to the thigh (31). However, in d-dimer research, the use of a single ultrasound as the gold standard creates the potential for bias. A gold standard that images both thigh and calf DVT, such as venography, will enable all patients with lower extremity DVT to be properly counted as having disease (Table 1A). However, to the extent that a gold standard such as duplex ultrasound fails to fully image the calf veins, patients with disease will be counted in the nondiseased category (Table 1B). This may lead to biased indexes of test efficacy. For example, if the true sensitivity of a d-dimer test for thigh DVT were higher than the sensitivity for calf DVT, the measured sensitivity would appear higher than it actually is. Most importantly for a "rule-out" test, the negative predictive value (NPV) will appear to be higher than it actually is because false-negative calf DVT results will be counted in addition to true negatives in the calculation of NPV (Table 1B). Both parts of Table 1 include actual data from a recent study (13) that suggest that these predictions are accurate. The purpose of this report is to verify the presence of this bias in d-dimer test research and determine its potential magnitude. To do so, we identified recently published d-dimer studies that provided data allowing comparison of the d-dimer test characteristics for overall lower extremity DVT (both thigh and calf) with those of thigh DVT alone.